The British Society of Gastroenterology - Gut

Gut, 1987, 28, A1328-A1412

The British Society of GastroenterologyThe Jubilee Meeting of the British Society of Gastroenterology was held in the University of London on15-18 September 1987 under the Presidency of Mr John Alexander-Williams. Below are printed the abstractsof the 364 oral and poster communications selected for presentation at the meeting.

INFLIAMMATORY BOWEL DISEASE

Enteral nutrition in severe colitis

S S C RAO, C [) HOLDSWORITH, ANI) A R W

FORREST (Gastroinitestinial U,iit and Depart-meait of Clinical Chemistry, Royal Hallam-shire Hospital, Sheffield) Small intestinalabsorption and tolerance of enteral nutri-tion was assessed prospectively in ninepatients with severe colitis. After admis-sion, the patients were starved for 24 hoursduring which stool output was collected.Thereafter 2 I/day of Fortison standardsolution was given enterally via a fine boretube for six days, after which patients ate anormal diet. During enteral feeding, a 5 g dxylose test and a five day faccal fat estima-tion were carried out. Stool output, symp-tom scores, daily weight, Hb, ESR andalbumin were also monitored throughoutthe 10 day study period and prednisolonesulphate was given intravenously. One hourblood xylose and five hour urinary xylosevalues were normal in all except one patientwith Crohn's disease. Fat absorption wasnormal in every patient. In spite of enteralfeeding, there was a significant reduction instool weight (p<0)(01) and stool frequency(p<0O001) over the 1i) day period. Therewas a signiticant improvement in symptomscores (p<0(0l), ESR (p<t)05) and per-centage weight loss (p<0.05). Serumalbumin improved [32 (28-37 g/l) v 28(23-36 g/l), median (range)l. All patientsachieved medical remission and one under-went an elective panproctocolectomy. Weconclude that even in severe colitis, smallintestinal absorption is normal, enteralfeeding is well tolerated and parenteralnutrition should not normally be necessary.

Diagnosing inflammatory bowel disease(IBD) with Tc 99m HPMAO labelledleucocytes

M E RODDIE, M PETERS, M J CARROI.I., J CALAM,

H J F HODGSON, AND P J lAVENDER (Depts ofRadiology and Medicine, Royal Post-graduate Medical School, Hammersmith

Hospital, London, and Amersham Inter-national, Bucks) "'Indium labelled leuco-cytes are an accurate means of detectinginflammation, particularly inflammatorybowel disease (IBD). Tc 99m as an alterna-tive radioisotope offers advantages of con-venience, lower radiation dose, andtheoretically higher image resolution. Anew Tc 99m chelate, HMPAO (hexamethylpropylene amine oxime) is a moderatelyefficient leucocyte label with selectivity forneutrophil leucocytes. We evaluatedabnormal scans after reinjections of auto-logous Tc 99m HMPAO leucocytes in 32subjects with proven or suspected IBD. Thenormal distribution of radioactivity issimilar to that seen with indium labelledcells, with additional activity in urine soonafter injection, and some bowel distributionin normal subjects after four hours.Occasional activity in the gall bladder indi-cates that this normal bowel activity followsbiliary excretion of free complex. Bowelactivity resulting from labelled neutrophilmigration to inflamed areas could be identi-fied much earlier than four hours, however,and as a result IBD could be confidentlydiagnosed. Twenty five of 32 patients hadabnormal scans showing inflamed bowel.The resolution of small bowel involvementis clearly superior to that with indium leuco-cytes. No false positives or false negativeswere encountered.

Use of cloned DNA probes to identifyCrohn's disease microbial isolates

J J MCFADDEN, J THOMPSON, E GREEN, S J

HAMPSON, J L STANFORD, AND J HERMON-TAYLOR (Department ofSurgery, St George'sHospital Medical School, and of MedicalMicrobiology, the Middlesex HospitalMedical School, London) Crohn's disease(CD) microbial isolates in longterm culturesfrom several laboratories have includedbacillary forms of mycobacteria, as well asthe scant growth of apparant spheroplastswhose identity has been uncertain. Conven-tional tests do not readily distinguishbetween some mycobacteria such as M

A 1328

paratuberculosis and M avium. Restrictedtotal DNA from a CD mycobacterialbacillary isolate was used to generate agenomic library in pGEM-1. Resultingspecific cloned DNA probes precisely dis-tinguished different mycobacteria on thebasis of restriction fragment length poly-morphisms (RFLP). A novel insertionalelement of about 1-5 kb (pMB22) wasidentified, repeated 1()-20) times through-out the mycobacterial genome. DNArestricted and probed with pMB22 gave amultiple banding pattern RFLP identical inthe DNA from each of three independentCD, and three Johne's disease M paraisolates, but not other mycobacterial DNAsuggesting a specific relationship betweenthis insertional element and pathogenicity.Similar studies on nanogram quantities ofDNA from one CD spheroplast isolateusing a novel cloned DNA probe for Mpara16S ribosomal RNA identified the sphero-plasts as mycobacterial but distinct from anyknown mycobacteria so far examined.Crohn's disease derived mycobacterialisolates appear heterogenous.

Adhesion of colitic E coli (EC) to isolatedhuman colonocytes

D A BURKE AND A T R AXON (GastroenterologyUnit, The General Infirmary, Leeds) Tissueculture and buccal epithelial cell adhesionassays have shown that E coli (EC) isolatedfrom patients with ulcerative colitis possessan adhesive property. In this study humancolonocytes have been used to assess in vitroadhesion. Colonocytes from mucosal biop-sies were isolated by gentle agitation in thepresence of hyaluronidase at 37°C. The testbacterial suspension was incubated with theisolated colonocytes in Ham FIO at 37°C for30 minutes. Adhesion was assessed underphase contrast and expressed as the numberof bacteria adherent to brush border (BB)and BB/basolateral (BL) adhesion ratio.The BB score of 12 colitic EC isolates tocolitic colon, mean=2()3, is significantlyhigher than control EC, mean=-028,p=(0.Ol. The BB/BL ratio of colitic EC,

on January 12, 2022 by guest. Protected by copyright.

http://gut.bmj.com

/G

ut: first published as 10.1136/gut.28.10.A1328 on 1 O

ctober 1987. Dow

nloaded from

http://gut.bmj.com/

7he British Society of Gastroenterology

mean=0-33 compared with a mean of 0-1for control EC is significantly higher,p (0001. Buccal epithelial cell adhesiveability correlates with both the BB adhesionindex and the BB/BL ratio, p=0)001. Thisstudy shows that colitic EC have a greaterability to adhere to human colitic colon thancontrols. The higher BB/BL ratio for coliticEC suggests that adhesion is specificallydirected to the brush border. These datasupport the view that EC may have a role inthe pathogenesis of this disease.

Delayed-release 5-aminosalicylic acid(5-ASA) and sulphasalazine (SSZ) in thetreatment of mild to moderate ulcerativecolitis (UC) relapse

S A RIl EY, V MAN!, M J GOODMAN, AND L A

I URNBER(G (University Departmenit ofMedicine, Hope Hospital, Salford; LeighInfirinary, atid Bury General Hospital,Manchester) Sulphasailazine is of benefit inUC relapse but side effects may limit its use.5-ASA appears less toxic. We have there-fore compared SSZ, equivalent dose 5-ASAand high dose 5-ASA in mild to moderateUC relapse.

Sixty one patients (32M:29F, aged 20-78years) were randomly allocated to eitherSSZ 2 g, 5-ASA 80X) mg or 5-ASA 2-4 g dailyin a double blind four week trial. Onepatient defaulted leaving 19 SSZ, 20 5-ASA80X) mg and 21 5-ASA 2-4 g for analysis.Groups were comparable for age, sex,cxtent of disease and pretrial SSZ intake.Six patients were withdrawn, four SSZ (twodeterioration, two side effects) and two5-ASA 2-4 g (deterioration).

Within treatment comparisons revealedsignificant improvement of (a) sigmoido-scopic grade in the SSZ group, (b) rectalbleeding and sigmoidoscopic grade in the5-ASA 800 mg group and (c) stool fre-quency, rectal bleeding and mucus andsigmoidoscopic grade in the 5-ASA 2-4 ggroup. Symptomatic remission occurred in2 SSZ, 300o 5-ASA 80X) mg and 42%5-ASA 2-4 g (p<0.05 v SSZ).

Greater improvement of rectal bleedingoccurred in the 5-ASA 2-4 g than SSZ group(p<(-(05). Side effects, however, weresimilar in the three groups.

High dose 5-ASA is more effective thanSSZ in the treatment of UC relapse.

Causes and prevention of postoperativefistulae in Crohn's disease

FRANCOISE IIEYEYN, M C WINS! LT, HILARY

ANDRE.WS, R N Al.lAN, M R B KEIGHI LY, AND

J ALEXANDER-WILI-IAMS (T/ic Gen1eralHospital, Birmingham) From a series of 59patients with enterocutaneous fistulae com-plicating Crohn's disease, 31 occurred in 23patients within a month of operation.Four were referred having presented the

fistula within 10 days of appendicectomy bysurgeons inexperimented in Crohn'sdisease. All were ileocutaneous, three hadresection and one a stricturoplasty; allresolved.

Suture line leakage accounted for 18fistulae and nine, presumed to be caused byoperative trauma, originated from a dis-sected area of gut not affected by Crohn'sdisease. Previous reports suggest that themajority of such fistulae will heal spontane-ously. In this series, however, only 12 of the27 resolved without further operation.The principal predisposing factors in

suture line leakage were adjacent septicfoci, hypoproteinaemia and colonic anasto-mosis. In some severely compromisedpatients a primary stoma with delayedanastomosis might avoid leakage.

In another series of patients having exten-sive dissections, occult operative traumahas been rendered overt by CO, distensionof the gut and underwater inspection (cycletyre puncture manoeuvre). All suchdetected bowel breaches have been suturedwithout subsequent fistulae. We suggestthat routine adoption of the insufflationmanoeuvre might have prevented five of thenine operative trauma fistulae in this series.

Strictureplasty for Crohn's disease - a nineyear experience

[F C B DEIIN, N J MCN MORIENSEN, M G W

KETI LEWELL, 1) P JEWE.LL, AND) THIL .AlAE E C G

LEIE (John Radcliffe Hospital, Oxford) Therole of strictureplasty (SP) is controversialin the management of obstructive Crohn'sdisease (CD). In Oxford, since 1978, 2(0patients have undergone 26 SP operationswith a median follow up of 30 (range 2-102)months. Crohn's disease had been presentfrom 0-240 (median 12) months. Sixteenpatients previously had resections for CD.Eighteen patients received high doseparenteral steroid therapy preoperatively.Strictures were confirmed before surgery bysmall bowel enema and identified opera-tively by intestinal intubation. The mediannumber of strictures identified at surgerywas two (range 1-17): 16 patients had oneSP, 10 patients between two and seven SP.The length of stricture plastied bowelranged from 1-17 (median 3) cm. Intestinalresections were performed concurrently in12 patients. Four patients subsequently

required a further 14 SP between 12 and 36(median 18) months after the initial SP. Allbut one of the previous SPs were patent. Nopostoperative complications arose from theSP sites. Intraperitoneal abscess developedin one patient from a duodenal leak. Ourfurther experience indicates that SP is aneffective and safe procedure which has aplace in minimal surgery of both active andinactive CD, provided no distal obstructinglesions are overlooked.

I IVER

Effect of hepatic artery embolisation onsurvival in carcinoid syndrome

M COUPE, A HEMMiN(iWAY, Ii J F HIODGSON,ANI) 1) J AL ISON (Depis of Medicine antd(Radiology, Royal Postgraduate MedicialSchool, Hammersmith Hospital, Lonidon)Hepatic artery embolisation relievessystemic hormonally mediated symptoms inthe carcinoid syndrome, and local symp-toms of hepatic pain. We documcnt here theeffect of this procedure on survival in aretrospective survey of 63 patients withcarcinoid syndrome seen in one centre over10 years. Twenty four underwent hepaticartery embolisation, and in a further sixembolisation was intended, but technicallynot feasible (portal vein obstruction orinadequate arterial access). The indicationfor embolisation was severe systemic symp-toms poorly controlled by pharmacologicaltherapy, or marked local pain. Ovcrall, inthose patients followed to death, the meansurvival from symptom onset was longer inthose undergoing hepatic artery embolisa-tion, but this was not statistically significant(58-7±9.7 (SEM) months n= IX v449+59-6,n= 1 1). A more accurate comparison can bedrawn between those patients successfullyembolised, and those in whom the pro-cedure was unsuccessful, comparing sur-vival from the angiographic procedurc.There was no difference in survival betweenthese two groups (embolised mean 18.0+3-26 months n= 18 v embolisation technic-ally impossible mean 19-6±6-5 monthsn=6). This study confirms that the hepaticartery embolisation does not prolong sur-vival in the carcinoid syndrome, despite itseffectiveness in relieving symptoms.

Cellular chemotaxis to bile after liver trans-plantation

A1329


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroenterology

[) Ii ADAMS, D BURNEFn, R A STOCKLEY, AND

E EI.IAS (Liver Unit, Queen ElizabethHospital, Edgbaston, Birmingham) Themechanisms of cell recruitment to liverallografts during rejection are poorly under-stood. Chemotactic factors may be impor-tant but have not been studied in man. Asthe cellular infiltrate is centred on bile ductswe have looked for chemotactic activity inserial bile samples from 13 liver transplantrecipients.

During nine episodes of gratft rejection,bile became chemotactic for both neutro-phils and mononuclear cells when com-pared with bile taken from seven stabletransplants (neutrophils rejection: range2 2-22, median 4*2xcontrol; stable: 0(44-2-7, median 1-73; p<0)t)05: monocytesrejection: 0-89-7-03, median 2 12; stable:0(55-2-63, median 1 08; p<0).025).

Bile taken two to four days before clinicalrejection was chemotactic for lymphocytes(2-3-8 9, median 3-2; p<0 (X)I) but duringclinical rejection inhibited chemotaxis(0( 144- 108, median 0(74, p<0.025) whencompared with bile from stable transplants(0.97-1-8, median 1.4).We conclude that (1) chemotactic factors

are important in cell recruitment to reject-ing liver grafts. (2) Lymphocytes areattracted to the graft before the onset ofclinical rejection. (3) During clinical rejec-tion lymphocyte chemotaxis is inhibitedwhereas other cells are attracted to thegraft.

Distribution of calcitonin gene relatedpeptide (CGRP) and substance P-containingnerves in liver: an immunohistochemicalstudy

A [) BURT, M GilL ON, E WISSE, J M P0l.AK, ANDR N M MaCSWEEN (University Department oJPathology, Western Infirmary, Glasgow,Lab for Cell Biology anid Histology, FreeUniversity ojf Brussel.s (VUB), Belgium,Department of Histochemistry, Royal Post-graduate Medical School, Lonldon) We haveinvestigated the distribution of nerve fibrescontaining the regulatory peptides, calci-tonin gene related peptide (CGRP) andsubstance P in guinea pig and rat livers.These peptides are closely associated withafferent neurones in the peripheral nervoussystem. The livers of Dunkin Hartleyguinea pigs and Wistar rats were perfusionfixed with 04%, parabenzoquinone. Tissueblocks were washed in 15%, sucrose andfrozen in liquid nitrogen. Polyclonal rabbitanti-CGRP and anti-substance P antibodieswere used at dilutions of 1:200 to 1:800 on

5 and 12 ,tm sections with an indirectimmunofluorescence technique. CGRP andsubstance P containing fibres were identi-fied within portal tracts in the livers of bothspecies. Fibres were most frequently seenaround hepatic artery branches but werealso seen in close proximity to bile ducts andportal vein branches. No intra-acinar fibreswere identified. There was abrogation ofstaining for both peptides in the livers of ratstreated in the neonatal period with thesensory neurotoxin capsaicin. These resultsindicate that the intrahepatic vasculaturebut not the liver parenchyma receives apeptidergic sensory nerve supply.

Management of increased cerebral bloodflow in patients with fulminant hepaticfailure

R J EDE, C D GOVE, AND K WILLIAMS (LiverUnit, Kings College Hospital and MedicalSchool, London) Raised intracranialpressure (ICP) is a major cause of mortalityin patients with fulminant hepatic failure(FHF) and is generally attributed to thedevelopment of cerebral oedema which is afrequent autopsy finding. It had been sug-gested recently that ICP might also beincreased as a result of dilatation of thecerebral blood vessels resulting in increasedcerebral blood flow (CBF). We thereforedetermined CBF using the iv '-"Xe clearancetechnique in 15 patients with FHF caused byparacetamol overdose (10), viral hepatitis(three), and halothane hepatitis (two).Patients were studied in grade fourencephalopathy after intubation andmechanical ventilation. In every case CBFwas increased with a mean value of 201ml/100 g brain/min (range 137-310) over aPaCO. range of 4-5-6-0 kPa (Cf normalvalue of < I 10 ml/min). Cerebral blood flowvalues did not differ significantly betweenthe three main aetiological groups.Cerebral autoregulation in response tochanges in PaCO. was assessed in sevenpatients: CBF increased linearly over thePaCO. range 2-5-7.5 kPa in all but onepatient in whom this autoregulatoryresponse was lost a few hours before death.Raised CBF was reduced by vigoroushyperventilation and more rapidly by ivAlthesin (we have shown that both thesemeasures reduce increased ICP in FHF).These data indicate that increased CBFcontributes to raised ICP in patients withFHF and have important therapeutic impli-cations for the management of patients withthis life-threatening complication.

Portal tract lymphocyte populations inprimary sclerosing cholangitis

J A SNOOK, K A Fl.RMING, A HERYET, P KELILY,D P JEWEI LL, AND R W CH{APMAN (Departmentof Gastroenterology and Nuffield Depart-ment of Pathology, John Radcliffe Hospital,Oxford) Recent evidence suggests thatimmunological mechanisms may be ofimportance in the pathogenesis of primarysclerosing cholangitis (PSC). We have pre-viously shown that patients with PSC havereduced numbers of CD8 cells in peri-pheral blood. Here, portal tract lymphocytesubsets have been analysed in cryostat sec-tions of liver biopsies using monoclonalantibodies and an immunoperoxidase tech-nique. Liver biopsy sections from ninepatients with precirrhotic PSC, ninepatients with primary biliary cirrhosis(PBC) and seven patients with histologic-ally normal livers were examined. The Tlymphocyte was the predominant portaltract mononuclear cell in all three groups.Mean total T lymphocyte counts per portaltract (+/-l SD) were increased in both PSC(173+/-105) and PBC (210+/-110) com-pared with controls (42+/-27). Similarincreases were seen for both CD4' andCD8 cells so that the CD4t :CD8' ratio was1-49 in PSC, 189 in PBC and 1 63 incontrols. In contrast, B lymphocyte countswere low in PSC (8+/- 14), PBC (9+/- 14)and controls ( 1+/- 1). These results suggestthat T cell mediated immune mechanismsmay play an important role in the patho-genesis of PSC. In contrast with the situationin peripheral blood, the portal tract CD4':CDSt ratio is similar in PBC and PSC.

Changes in the hepatic haemodynamics withthe development of liver metastases

D M NOi-l, S A JENKINS, S GRIME, J YAIES, D WDAY, J N BAXTER, AND I G COOKE (UniversityDepartmenits of Surgery, Pathology anidNuclear Medicine, Royal LiverpoolHospital, Liverpool) The ratio of thehepatic artery flow (HAF) to total liverblood flow measured by dynamic scinti-graphy, the hepatic perfusion index (HPI),is reported to predict the presence of occultliver metastases in patients with colorectalcarcinoma. The aim of this study was todetermine the mechanisms responsible forthe alteration in the HPI with the develop-ment of micrometastases. Metastases wereinduced in rats by an intraportal injection of8x 10( Walker carcinosarcoma cells.Animals were studied at two, four, and sixdays post inoculation. The HPI was

A 1 330)


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


measured by dynamic scintigraphy andhepatic artery flow (HAF) and portalvenous inflow (PVI) using '7Co micro-spheres. Intrahepatic arteriovenous shunt-ing (HS) was determined after an hepaticartery injection of FTc"' labelled macro-aggregated albumen (4t) u). Micrometa-stases were present at four and six days postinoculation. The HPI at four days (51-80+1-68; n=16) and at six days (67-81±3-38;n= 16) was significantly greater (p<0-001,Mann Whitney) than in normal rats (33-78±1-77; n=20). Hepatic artery flow did notchange with the development of metastasesbut PVI was significantly decreased at sixdays (p<0-0)1) and was associated with asignificant increase in splanchnic vascular(SVR) and portal venous resistance (p<0-001). At four and six days after inocula-tion HS was also significantly increased(p<0(005). These results indicate (1) theHPI alters with the development of micro-metastases secondary to decreased PVI, (2)the reduction in PVI is related to a risein hepatic vascular resistance consequentupon increased intrahepatic shunting.

BASIC SCIENCE POSTERS

Penicillamine and cysteine inhibit acetalde-hyde induced superoxide production by theneutrophil

A J K WIlTLIAMS AND R E BARRY (Department ofMedicine, Bristol Royal Infirmary, Bristol)Acetaldehyde binds non-enzymatically toliver membrane proteins. We have shownthat penicillamine and cysteine will inhibitthis binding, and that the product of thisbinding will activate complement andstimulatc the neutrophil both potentialmechanisms of cytotoxicity in alcoholichepatitis.We have studied the effect of penicilla-

mine and cysteine in modifying neutrophilsuperoxide release in response to acetalde-hyde altered rat hepatocyte membranes.Liver membranes were preincubated inI mM acetaldehyde+equimolar or excess(5/1 molar ratio) penicillamine or cysteine,pH 7-4, two hours, 18'C; washed and thenexposed to neutrophils with superoxiderelease measured by cytochrome c reduc-tion in the presence or absence of super-oxide dismutase.

Pre-exposure of the membranes toequimolar penicillamine and acetaldehydesignificantly reduced superoxide releasecompared to acetaldehyde alone exposed

membranes from 154±16 (mean±SEM) to77+ 10 nmol O0/10' cells/min n=9, p<0 01.Pre-exposure to equimolai cysteine signifi-cantly reduced superoxide release to94± 12, n=9, p<0-01. A further reductionoccurred when in molar excess (penicilla-mine 33±7, cysteine 57±13, n=9). Theseagents could have potential benefit in limit-ing neutrophil stimulation in alcoholichepatitis.

Measurement of pepsin synthesis in manusing 1-'3C-leucine

M E CORBETT, E J S BOYD, J G PENSIFON, K G

WORMSILFY, AND M J RENNIE (Departmnents ofPhysiology and Therapeutics, DunideeUniversity, Dundee, Scotland) We havestudied the effects of increasing doses ofpentagastrin (Pg) on gastric secretion ofacid, pepsin and total protein. In addition,we have used the rate of incorporationof I-'`C-leucine into the YCA-insolublefraction of the gastric aspirate as an index ofpepsin synthesis.On separate days 10 subjects received Pg

0, 0(25, 0.5, 1-0, 2-0, or 4-0 )tg/kg/htogether with a primed infusion of '3C-leucine 0-25 or t)*5 mg/kg followed by 0-25or 0(5 mg/kg/h respectively. Infusions werecontinued for 210 min. Gastric juice wasaspirated in 30 min batches. 'C enrichmentof plasma keto-iso-caproate was measuredevery 30 min. Affinity chromatography con-firmed that "3C-leucine was incorporatedsolely into pepsin.The dose of Pg required for half maximal

acid secretion was 0-13 tg/kg/h, and for halfmaximal pepsin secretion was t) 10 [tg/kg/h.Incorporation of '-C-leucine into the TCAprecipitate was linear with time, and therate was proportional to the plateau 'ACenrichment of plasma keto-iso-caproate. Pgdid not significantly alter from basal the rateof incorporation of '-3C-leucine into pepsin.These findings suggest that increased pepsinsecretion in response to increasing doses ofPg is not caused by increased pepsin synthe-sis, but is attributable to recruitment of cellshaving similar synthetic activity.

Gastric acid regulates postprandial somato-statin release in man

M R LUCEY, P D FAIRCLOUGH, AN[) J A Hi WASS

(St Bartholomews Hospital, London) Therole of gastric acidity in mediating release ofsomatostatin like immunoreactivity (SLI) inresponse to orally ingested nutrients in manis uncertain. We studied in healthy men the

effect of cimetidine (1()t) mg/h iv for fourhours) on circulating SLI levels basally anidwith ingestion of a mixed meal. In somestudies, intragastric pH was measuredradiotelemetrically and either O)-1 N HCI or0(9%/o NaCI infused intragastrically. Infurther studies, an intraduodenal infusionof fat was given with/without iv cimetidine.

Cimetidine (C) reduced meal stimulatedSLI levels: meal alone: 83± I)0% min, meanintegrated increment from basall, ±SE;meal+C: 51±12% min, p<0()-5 (n=12).Intragastric infusion of HCI ameliorated theeffect of C on mean postprandial gastric pH:meal alone: 2-21; meal+C: 3-79; meal+C+HCI: 3-15; and prevented the cimetidine-induced reduction in postprandial Sl Ilevels. Meal alone: 75+ 18% min, meal+C:28±15% min, meal+C+HCI: 65±12'Y0 min(n=6). Cimetidine did not affect the rise inSLI levels during intraduodenal fat infusion(n=5). These data suggest that (1) cimeti-dine reduces postprandial SLI secretionindirectly by altering intragastric acidity,(2) intragastric acid is a factor regulatingnutrient stimulated SLI levels in man.

Fat is a luminal stimulus for gastric somato-statin release

M R LUCEY, J DELVAILIE, AND YAMA[)A

(University of Michigan Medical C'enter,Ann Arbor, Michigan, USA) The mecha-nisms regulating nutrient stimulated releaseof somatostatin like immunoreactivity(SLI) from the gut are unclear. We investi-gated the effects of fat and its constituentsadministered intra-arterially or intralumin-ally on SLI release from the isolatedperfused rat stomach. Fat emulsion (micro-lipid), oleic acid and glycerol were adminis-tered intra-arterially in three dilutions(3x 10( 5, 3x 1) 4, 3x 10 ') in buffered per-fusate. Isoprenalin 10 SM was infused intra-arterially at the end of each experiment, aresponse exceeding 20)0% of basal levelsindicating a viable preparation. In intra-luminal studies, oleic acid and glycerol,each diluted 10-fold in 0.9%0 NaCI andundiluted fat adjusted to pH 6, were admin-istered for three consecutive 10 minuteperiods. Neither intra-arterial fat, oleic acidnor glycerol stimulated SLI release at anydose. Intraluminally administered fatincreased SLI release in a time dependentfashion reaching significance in the 3rdperiod. Period 1: 115±15% of basal(mean±SE), period 2: 116± 1)%, period 3:140)±12%, p<0-05, n=7. There was noincrement in SLI response to either oleicacid or glycerol intraluminally. These data

A 1 33 1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroetterology

suggest that (1) fat stimulates SLI releasefrom the isolated perfused rat stomach,(2) this action is dependent on presentationof the stimulus to the lumen, (3) stimulationof SLI release by fat does not depend on itsmetabolism or absorption.

Mechanisms governing the biphasic patternof gastric emptying after truncal vagotomyand pyloroplasty

N PARR, S GRIMtL, M CRITC iII IFY, J BAXTIR, AND

C MAC KI I (University Department of Surgerxi

anid Depairtmnetnt oJ'Nuclear Medicinie, RoyalLiverpool Hospital, Liverpool) Aftertruncal vagotomy and pyloroplasty (TV+ P)gastric emptying (GE) of liquids is biphasic,with rapid early GE often producing post-vagotomy symptoms. Factors restrictingthis phase and provoking transition to thesecond, slow phase are unknown. Weinvestigated the contributions of smallbowel resistances, osmoreceptor feed backand sympathetic inhibitory reflexes to smallbowel distension. Sixty GE studies were

done, using gamma camera imaging ofradiolabelled 15'Yo dextrose, on six dogswith TV+P and fitted with proximal duo-denal cannulae. With the cannula closedGE (%o) was initially rapid, followed bystasis (15 min 32±5-3, 60 min 34±4X8,mean±SEM). Opening the cannula alone,produced precipitous GE (76±4-2p<000 1, 88±2-6 p<0001 ANOVA). Withdistal duodenal instillation of 15% dextroseat a rate equivalent to GE with the cannulaclosed, GE remained faster than in theinitial studies (50±7-0 NS, 65+6-8 p<0-01),but was slower than with diversion alone(p<0-05). Normal saline instillation did notdelay GE (78+ 10-6, 90+5.4). Finally, duo-denal instillation of 15% dextrose prior toingestion of the test meal caused slowerinitial emptying than in cannula closedstudies, without subsequent stasis (24±4-5,47± 10(6). These results indicate that afterTV+P, small bowel resistances play a

significant role in controlling GE. Osmo-receptor responses persist after TV+P, butsympathetic inhibitory responses are notinvoked.

Gastrin - in vitro and in vivo studies ingastric and colorectal cancer

S WATSON, 1. DURRANT, D 1. MORRIS, AND

CG GCEROUILAKOS (Departments of Surgery and(Cancer Research, University Hospital,Nottingham) Gastrin is trophic for bothgastric and colorectal carcinomas in animal

models but in vitro studies have been lessreproducible. The recent innovation of syn-chronisation of cell cultures with thymidinewas used to study nine gastric or colorectalcell lines; no stimulation was seen withpentagastrin or G17. Using the synchro-nised assay, however, two cell lines showedan increase in growth rate (selenomethioneincorporation) to G17 10 [tg/l of 110 and120%. The most responsive of the cell lineswas xenotransplanted into mice and half theanimals treated with G17 10 ug/day. Not allxenografts grew but of the five fastest grow-ing tumours in each group the diameter ofthe tumours was significantly greater in thetreated group by 17 days (p<0(0l). Whencells from this xenograft were cultured invitro a 148% increase in growth was seen.We also studied the in vitro growth offreshly disaggregated human tumour cellsand showed a response to pentagastrin inone gastric and -/4 colorectal tumours. Thisresponse was retained at in vitro passage 2but lost by passage 6. Gastric and colorectaltumours are often sensitive to gastrin in vivo- this stimulation is often lost or reduced inin vitro cultures.

Changes in mucus glycoprotein biosynthesisat the tumour site in gastric cancer

N K DHIR, R L SIDEBOI IIAM, J SPENCER, AND

TIIE LATE J SCHIRAGER (Department ofSurgery, Roylal Postgraduate MedicalSchool Hammersmith Hospital, Doi CaneRoad, London) Immunochemical investi-gations and studies of cultured tumour cellshave shown that carbohydrate biosynthesisof mucus glycoproteins is altered in gastriccarcinogenesis. To extend these observa-tions we have examined the structures ofmucus glycopolypeptides from non-malignant mucosa (20 specimens; indi-viduals with benign peptic ulcer, or diseasefree), and from uninvolved mucosa (33specimens) or tumour site (32 specimens) inpatients with stomach cancer. Carbohy-drates and amino acids were measured bygas chromatography and autoanalysis. Fromthis, the weight of carbohydrate associatedwith representative polypeptide core seg-ments of equal length was determined.A mean difference of 29%o (p<0.0001)occurred between glycopolypeptides fromuninvolved mucosa and tumour site. Thisresulted from a reduction in numbers of (1)carbohydrate chains, from a mean of 343 to276 chains/l100 amino acid units (p=0-0( ),associated with a decrease of serine andthreonine residues in the polypeptide coreand (2) structural units per carbohydratc

chain, from a mean of 6-85 to 6 (p=0 01),attended by a decrease of fucosyl groups,from a mean of 1-6 to 1-05 per chain(p=(-02) in glycopolypeptides from thetumour site. Differences between mucusglycopolypeptides from non-malignant anduninvolved mucosa were small, and notstatistically significant. These data showthat carbohydrate biosynthesis is curtailed,and that biosynthesis of the polypeptidecore may be abnormal, in mucus glyco-proteins produced at the tumour site ingastric cancer.

N-methyl-N '-nitro-N-nitrosoguanidine(MNNG) induces changes in rat and humanpepsinogen phenotypes

J DEFIZE, J K DERODRA, AND R H HUNT

(McMaster Untiversity, 1200 Maini Street W,Hamilton, Ontario, Canada) A character-istic human pepsinogen (PG) phenotypehas been described in association withgastric cancer and MNNG is known toproduce gastric tumours in the rat. Wedescribe changes in rat and human PGphenotypes, induced by MNNG in in vivorat experiments and in in vitro cultures ofboth rat and human chief cells. Rats werefed MNNG or placebo for five differenttime periods. PG phenotypes and histo-logical status of the mucosa were deter-mined from open biopsies taken underanaesthesia before, during and afterMNNG. The fastest electrophoretic banddecreased or disappeared as early as threeweeks after MNNG. The changes, observedin 17 of 32 rats, were permanent and con-sistently associated with gastric tumourlesions observed 10 months later (17 of 17).Chief cells were cultured in the absence orpresence of MNNG, which induced similarPG phenotype changes in both rat andhuman chief cclls after 7-1) days. In humanchief cells, a decrease of the Pg3 band,which is consistent with the 'carcinogenic'phenotype, was observed in two of sixpreparations, treated with MNNG.We conclude that MNNG induced

phenotype changes reflect mutations in thepepsinogen genes. The changes preceedgastric tumours and should be investigatedas a marker for the early onset of gastriccancer.

Lactulose/mannitol: an ideal screening testfor coeliac disease?

1. D JUBY, J ROTHWELL, AND A T R AXON

(Castroeenterology, Unit, The General

A 1 332


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

Tue Britisli Society of GastroenterologyA

InfirmarY, Leeds) Changes in intestinalpermeability are shown by differential sugarabsorption tests in coeliac disease, but arenot ideal for screening as the sugars arcoften difficult to analyse and chromato-graphy is usually required. We describe achemical technique for analysis of lactuloseand compare the lactulose/mannitol(La/Ma) test to a standard cellobiose/mannitol (Ce/Ma) test in two groups: (1)newly presenting coeliacs (17); (2) healthycontrols (12). The test solution comprises5 g lactulose, 2 g mannitol, and 22-3 gglucose in 1)() ml water (hyperosmolalityenhances discrimination between coeliacsand normals). This is ingested after a sixhour fast and urine collected for five hours.Lactulose is measured by an enzymaticmethod anid mannitol by spectrophoto-metry, the result being expressed as a ratiowhich eliminates extraneous factors (gastricand bladder emptying, etc). The medianLa/Ma test was )-08() (range 0)-)2-1-356) incoeliacs and 0(016 (range 0004-( (028) incontrols (p<()-()0l) as compared with amedian of )- 112 in cocliacs (range 0-024-0-720) and 0)-1 in controls (range 0-008-0-024) for the Ce/Ma test (p<0-00l). Thereis no significant difference between the twotests. This method fulfils the criteria for anideal screening test, it is of low cost, simpleto do and analyse, and is non-invasive. Itcompares favourably with the Ce/Ma testwhich is highly sensitive and moderatelyspeclfic.

Developmental changes in the villous uptakeor iron and enterocyte iron binding proteinsin the guinea pig duodenum

S K S SRAI, E DI.BNAM, ANI) 0 EPSTEIN (RoyalFree Hospital Sclool of' Medicinie, Lonidonl)In ai previous int vivo study we showed thatthe rate of iron uptake by neonatal guineapig duodenum is substantially higher thanthat of adults. We proposed that the failureof postnatal adaptation in man might be afactor in the pathogenesis of haemochroma-tosis. This study reports the use of auto-radiography to define changes in villouslocalisation of iron uptake and gel filtrationof enterocyte soluble supernatant (SSN) tostudy the ontogeny of iron binding proteinsin the guinea pig duodenum.

Autoradiography revealed that whilst allneonatal enterocytes took up "Fe, onlythose in the top half of adult villi take up theisotope.

Gel filtration of SSN on Sepharose 6Bindicate that in the newborn guinea pig themiajor "Fe-binding peak has a low molecu-

lair weight (MW< l()12)000) accounting tor70-800 of the eluted counts. In contraist,the major 'Fe-binding peaIk in adults wxas ahigh molecul'ar weight (MW>¢45000() poss-ibly ferritin) accounting for 62%S ot elutedcounts. The differences in iron bindingproteins and iron uptake along the villusmight aiccount tor the difference bctweenneonatial and adult iron absorption des-cribed in vivo. Similar studies in haemo-chromatosis might shed light on its patho-geniesis.

Localisation of VIP binding sites in humanand guinea pig gut using in vitro autoradio-graphy

R F POWIER, A F' BISHOP, M A (GIIAT I, S R BOOM,ANI) J M P0l1AK (Departmnent.s of' Histo-cheinistrv and Medicinie, RPMS, Hatmmer-smith Hospital, London) Despite numerousstudies of peptide receptors by bindingtechniques on isolated membrane preparai-tions, little information is available on theirprecise localisation. In this study, we haveused the tcchnique of in vitro autoradio-graphy to localise receptors for vasoactiveintestinatl peptide (VIP), one of the mostabundant peptides present in the entericnervous system, in humain and guinea pigsmall rind lairge intcstine, where it is knownto have potent aictions. Unfixed cryostiatsections werc incubated with radiolabclled('"I) ligand at a concentration of I nM.Specificity ot binding was confirmed byincubation of further sections with radio-labelled ligand and unlabelled VIP at aconcentration of I uM. Autoradiogramswere generaited by exposure of the sectionsto LKB-Ultrofilm or emulsion coated glasscoverslips. Arcas of dense binding could beidentified, correlating with the knownactions of the peptide. For cxampre, bind-ing sites appe(ared to be dense in the smoothmuscic and mucosa, in keeping with thepcptidc's effects as a muscle relaxant andstimulator of watcr and electrolyte secre-tion. The methods described here open newopportunities for study of the anatomicaldistribution of peptide binding sites in thegut. They may be particularly useful indefining the pathogencsis of certaindiseases.

Benzodiazepines stimulate duodenalepithelial bicarbonate secretion

J R IIHIYLINGS, S E IIAMPSON. ANI) A (ARNi.R(Bioscience D)ept, ICI P/harmnaceuticals

l)ivision, Al(lderlev P(ark, /Vlacclcsfield,(Cles/hire) Gastrointestinal protectivceffects of henzodiazepines (H3Z) in vivo aregenerally assumed to reflect a centralaction. We examined the influeLIce ofseveral BZ on duodenal HCo) secretion, aprocess which protccts the tissue f'romluminal H tby generating a standing pHgradicnt at the mucosal surface. Bull frogproximal duodenum wats stripped of sero-muscle layers, mountcd in a chamber andalkalinisation of the unbiuffered luminalbathing solution mea;sured by conitinuoustitration at pH 74 dilazeparm (1-01()( !Mserosill side) induced a dose-dependenit risein luminall alkalinisation and transepithelialpotential differenice (PD). Although somiie50)0-times less potent thian prostaglandin E,the increase in secretory rate was compar-able to the PG-induced tissue miaximia.Pretreatment with indomethacin ( I IM)had no eftect whereas 2,4-dinitrophecrolabolished the response to broth agonists.Dose-related stimulation of alkaline secre-tion and PD was aIlso observed with chlor-diazcpoxidc, nitrazcparm, flunitrazepamnand beta-ciarboline which increaised secre-tion by,70%, 60%, 90%, aind 55") respect-ively (aIll n=4) of the PGE, maximalresponse. These data show that BZ cian aictlocally to stimulate electrogenic, energydependent duodenal HCO( secretion by amechanism which is indepeindent of ciildo-genous PG production.

Preparation of osmotically active brushborder membrane vesicles from biopsysamples of human small intestine

S P SHIRAZI-B1i.ECIIFY, A G DAVIES, ANI) R B

BHIEt liii.Y (introduced by R M (ASE) (Dept ofBiochemistrv, Universits' College of Wales,Aberv.stwvtth, Dvfied, atidl Brontgl(at.sHospital, A bervystwyvt/, Dyied) The strategyhas been to prepare aind characterise indetail the brush border vesiclcs from differ-ent regions of the rabbit gut. This has beenperformed on a large scale. Brush bordervesicles were thenl prepared from hiopsysized portions of the rabbit gut. The proper-tics of these vesicles were compared withthose prepared on the large scale. No majordifferences were noted.We have used the technilques that were

developed with rabbit tissue, to preparebrush border vesicles from human jejunumand duodenum biopsy material. Thesevesicles have a characteristically high levelo alminopeptidase N. They also havc a Na'.-dependent D-glucosc transport systcmlocated within their membrane. This has a

A 1 333


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society ofG(lastroenlterology'

very high activity, 1 6-3-4 nmol 1)-glucosctransported/milin mg protein. Thc transportis inhibited by phlorizin. The time coursefor the Na '-depcndent uptake of D-glucoseby thcsc brush-bordcr vesicles shows atransieint accuniuliltion, that is, typicil over-shoot behaviour. Ihc level of accumulationof D-glucosc shows that the trainsportcr inhuman duodceunum aind jejununii is particu-lItrly active and that the movement of' D-glucosc through the membrane by passive.non-carricr mediaitcd mcchainisimis is vcrylow. These vesicles will entable the directinvestigation of transport mechanisms inhuman intestinal tissucs.

Induction of enteropathy by activated T cellsin human small intestine

T MAC()ONAI.I) ANI) Jo SPI.NCI'R (Paediatric(Gas.troeniterology', ,St Bartholomiew 'sHospital a11( l)epartment of Histo-pathology, University College, London) Tcells in the lamina propria of explants of16-22 week-old fetal human small intestinein organ culture were aictivatted by aiddingpokeweed mitogren (PWM) to the culturemcdium. This resulted in the appearance ofaictivated T cells (Il -2 receptor positive) inthe latmina propriai (but not the epitheliuni)and lymphokine secretioni into the organculture superiatant. Morphologically,explants stimulaited with PWM developedvillous atrophy atnd crypt hypertrophy after72 hours in culture. '[here was also adramatic increase in the number ot dividingcells in the crypts, demonstrated immuno-chemicatlly using the monoclonal antibodyKi67. Pokeweed mitogen only inducedthese changes in explants of Ietail smaillintestine which contained T cells. Finally,the effects of PWM were inhibited by cyclo-sporin A. These results unequivocably showthat activated I cells can produce a coeliac-like' enteropiathy in human smaill intestineaind provide a model in which the mecha-nisms important in the development ofenteropathy can be investigated.

Effect of rotavirus infection on water trans-port in infant mouse intestine in vitro

W (i SI'ARKEY, J (COiLLINS, K J WOR'I'ON, T X

WALI.IS., , J ClARKI., A J SPENCER, S J IIAI)I)ON.J STPTiH.N, M P OSBORNI, ANID ) C A CANDY(Institute of' (hild Health, Departmnentoj Microbiology an !Department oJPhysiology, Universiti of' Birnmitnghattm,Birmingham) An ini vitro intestinall perfu-sion system was validated in order to study

w,atcr and solute transport in infatnt mouseintestine, using "4C-polyethylene glycol as anon-absorbablc water marker. Controlsmall intestinc, perfused with WHO oralrehydration solution (Na 90 mM, K 20 mM,HCO) 30( mM) with glucose replaced withmcannitol, cxhibited a mean (SEM; n)steady-state net water aibsorption over aperiod ot 60 min of + 10-6 rd/cm/h (1-9;n=6). Absorption was inhibited ait 4° tomcan nct sccrction of -0-7 fd/cm/h (0)7;n=8). 'I'hcophylline induced a mean netw.ater secretion of - I X fld/cm/h (0(5; n=6).Meain recovcry of' 14'W-PEG was 97%. NetwCater secrCtioi wtas induced in mid-smallintestines by rotavirus infection of 1t) dayold mice. Secretory rate was maximal at-5-3 rt/cimi/h ( 1-4; n=7) 72 h postinfection.coinciding with the maximum severity ofsigns of diarrhoea anld vacuolattion of villoustip cntcrocytes. Peak EL1ISA titrc of rota-virus antigen in intestinail tissuc occurred at48 h. Water secretion was reversed to netabsorptioii (+5*4 dl/cm/h (I 1; n=6p<0-)0005) by perfusion with WHO oralrehydrationi solutioni containing 111 mMglucose. This is in accordance with theclinical success of WHO oral rehydrationsolution in humans with rotavirus diarr-hoea. Water absorption in the upper smallintestine and colon was not significantlyaltered by rotaviruIS infection.

Inhibition of ethanol induced leucotrienesynthesis and damage in the rat gastricmucosa by BW755C

N K BOUGHTiON-SMITIH ANt) B J R WHITITE(Department oJ Mediator Phartnaccologv,Welltcome Research Laboratories, Beckeni-harm, Kenlt) Oral administration of ethanolinduces damage and increases pro-ulcerogenic leucotriene (LT) release fromthe rat gastric mucosa. Using the lipoxygen-ase inhibitor, BW755C, the relationshipbetween LT release and ethanol induceddamage was further investigated.

Fasted rats (male, 180-200 g) were pre-treated with BW755C (5-50 mg/kg po) orvehicle, 30) min before oral absolute ethanol(I ml). After 5 min, the gastric damage wasassessed planimetrically and the mucosalformation of cicosanoids, tfrom choppedmucosa incubated for 20 min at 37C, deter-mined by specific radioimmunoassays.

Immunoreactive LTB4 and LTC4 releaseby the control gastric mucosa (5-1 ±0-5 ng/gand 81+27 ng/g of tissue respectively,mean±SEM n=5) was increased afterethanol (to 21+ 2 ng/g and 164±14 ng/grespectively. n=5; p<0).05). Ethanol

induced damage, which involved 3()+4 YO(n=6) of the total mucosal area was reduced(p<0-05) by BW755C at doses of 20) mg/kganid St) mg/kg (42±13%) and 63± 12%.inhibition respectively), as was the mucosalformation of LTBI (79+4%O and 78±7inhibition respectively, p<0-0l) and LTC 4(92+2 MO inhibition at 20 mg/kg, p<t)-0).The formation of 6-keto-PGFI,, wasunaffected by BW755C.BW755C thus inhibits both ethaniol

induced gastric mucosal damage and theincrecased formation of LTBI and LTC4, butdid not affect formation of the mucosalprostanoids. The failure of BW755C toprevent completely the ethanol induceddaimagc when there was near-mi-aximalinhibition of LT synthesis may indicate thaitadditional mechanisms are involved in itsprotective action.

Acute intestinal ischaemia: difTerentialresponse of gut neuropeptides

1. Ml.l-.AGiROS, J S (iGllA., P K MiL D)iFRRY,J D)OMIN, M A GIHATIEI, ANI) S R BiOOM (D)epart-ttinelt of Medicinte, Royal PostgraduateMedical School, Hammsleerstnitlh Hospiial,Dii C(tne Road, Loiidono) Superiormesenteric artery (SMA) occlusion is a lifethrealtening condition associated with colin-plex pathophysiologicill disturbanrceslocally and systemicilly. We hatve investi-galted the response of certain gut neuro-peptides in experimentail SMA occlusioii inthe acnaesthetised rat. The SMA waisclaimped at its origin for five (n 10)) or 2t)(n= 10) minutes. After release of the clamnp,blood sa mples were talken by aorticpuncturc and all the areas of the gut wereextracted in boiling 0-5 M aicetic acid.Neuropeptides in plasmat and tissues weremeasured by specific radioimmunoassays.Results, analysed by ANOVA, revealed arise in plasma calcitonin gene relatedpeptide (CGRP): controls (n =i)) 65±1(,five minute ischaemia l1)6± 16 (NS), 20minute 147±36 (p<0(05) (mean±SEMpmol/l) but not in plasma neuropeptide Y(NPY). In the gut jejunal CGRP fell from31-1± 16 to 238± 1.5 (p<()-()5) and to22 1 + 1 (p<0-0l) and substance P froni70±5' to 46±2 (p<0).()l) and to 51+3(p<t).05) after 5 and 20 minutes ofischaemina (mean ±SEM pmol/g) reflectinglocall release. Ileal and right colon conceni-traitions fell but not significantly. In contralstNPY and bombesin concentrations wereunaltered. Therefore in acute ischaemil thevatsodilatory neuropeptides CGRP and sub-staince P. but not the vatsoconstrictors NPY

A 1 334


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


and bombesin, are released locally suggest-ing possible mechanisms that may operateto limit the extent of ischaemic damage.

Electrophoretic analysis of microvillusmembranes in coeliac disease

P W PEMBER ION, R W l OBI LY, AND R 1101 MI.S(Unii'ersitvi Department of CaGastro-enterologi', Manchester Roval Infirmary,O.Vfiord Road, Manchester) We have pre-viously described a technique for analysingmicrogram amounts of intestinal micro-villus membranes by two dimensionalelectrophoresis (protein mapping) and usedit to determine the protein composition ofnormal human membranes. We have nowapplied the tcchnique to the analysis of themembrane in coeliac disease.

Jejunial biopsies were obtained from con-trols, and from coeliac patients in rclapseand remission. Microvillus membraneswere isolated by the Ca` precipitationmethod, solubilised in 20O SDS and radio-labelled with '4C by reductive methylationbefore electrophoresis. One dimensionalseparation was performed by conventionalS[)S-polyacrylamide gel electrophoresiswhile two-dimensional analysis was by iso-clectric focussing followed by SDS-PAGE.Proteins wcre detected by fluorography ofthe dried gels and glycoproteins by lectin-affinity staining of Western blots.Compared with controls, the protein map

in untreated coeliac disease was grosslyabnormal. Numerous high Mr proteins,including the brush border enzymes andother glycoproteins were absent, whilesmaller proteins - for example, actin, ,wereunaffccted. After treatment, the proteinniap returned essentially to normal withrestoration of the high Mr proteins, and nospecific abnormality of the coeliac micro-villus membrane could be identified.

Acetate uptake in rat small intestine is notsodium dependent

A J M WATSON, F A BRFNNAN, M J G FARTIIING,ANI) P 1) FAIRCLIrOUG-H (Department of Castro-enterology, St Bartholomew's Hospital,London1) We have shown that acetatestimulates sodium absorption in rat andhuman small intestine. The aim of thepresent study was to establish whether Na-acetate cotransport can account for thiseffect. Acetate uptake was studied in brushborder membrane vesicles (BBMV) whichwere prepared using a Ca-precipitationtechnique and compared with glucose

uptake which is known to be via a Na-glucose cotransporter. It was found thatglucose uptakc was ( 1) stimulated 35-fold inthe presence of 10)() mM Na compared withK. (2) a further two-fold stimulaition occur-red when a negative diffusion potential wascreated by substituting thiocyanate forchloride and (3) varying the osmolality ofthe medium altered the equilibrium uptakeof glucosc. In contrast, no difference wasfound in acetate uptake (1 mM) in thepresence of Nai or K. Uptake viaried linearlywith time up to two hours and there was noevidence of saturation of uptake at higherconcentrations up to 10)0 mM. Uptake watsenhanced at pH 5-5 compared with pH 7-4(672.5+65 v 438 7+23.1 pM/mg protcin.respectively; p<0t)*01). Acetate could notcreatc a negaitive diflusion potential forglucose uptakc. Vairying the osmolality didalter acetate uptake at 90 min. These datalend no support for the presence olf a Na-acetate co-transporter but rather suggestnon-ionic diffusion of unionised acetate, themajority of which rcmains in the BBMVmembrane in this system.

Denervation is the prime cause of reducedwater absorption in successful rat smallbowel transplants

A J M WATSON, I' A lEAR, E .1 IL.[II01-, R F M

WOOD, AND M J G FARTHING (Depts ofCGastro-enterology and Suirgery, St Bartholomew'sHospital, London) We have previouslyshown that in successful rat small boweltransplants impaired water absorption can-not be accounted for by mucosal damage.To determine the cause of this reductionwater absorption was studied in allografts(n= 1)) and compared with (1) Thiry-Vella(TV) loops (n= 1)), (2) isografts (n=6),(3) denervated TV loops (n= 5), and (4)ischaemic TV loops (n=4) using a steadystate perfusion technique of isotonic 30 mMglucose-saline nine days after transplanta-tion. All animals were treated with anidentical dose of cyclosporin A (15 mg/kgfor seven days only). Water absorption wasreduced by 55%, in allografts compared withTV loop controls. This was not hecause ofrejection as there was a similar reduction inisografts (median 42-6, range (17 5-77-7) vallografts (median 39t), range (19.6-54-2)dI/min/g, respectively). When TV loopswere denervated, however, water absorp-tion was reduced to levels found in allo-grafts (median 26.0, range (8X7-48t)).Water absorption was not reduced inischaemic TV loops. In contrast, no differ-ence was found in glucose absorption

between allografts and the other controlgroups stutdied, suggestillg vilIlLs absorptivefunctioni remains intaict. We concluide thatthe rectioll of water absorption inSuccCssfIul sm1a1ll bowel transpaiits isprinarily caLused by denervation rather thanrejection or ischaemnia.

Segmental variability of electrogenic Na+transport in human colon

( I SANi)IDE ANI) : MC(iLONF. (l)eparlteni of

Medicine, Hope Ho.spial (Universitv ofManchester School of' Medicine), Sallford)Electrogenic Na' transport is present innormal humran distal colon and is enhancedby aldosteronie. 'I'he Na' chalnnel blockermiloridc markedly inhihbits hasal electro-

genic Na' trainsport in humain disital colonibut has considerably less effect in huLimianproximal colon. To further study the baisisfor this segmental difterence in amiloridesensitivity, the effects of aimiloride ( I) M)and Na' replaicenient with choline weredetermined in proximall (asccnding) anddistal (sigmoid/proxim-al rectumiii) coloniresected from patients with cancer ordiverticular disease. In NaCl Ringer, basalelectricail properties of proxinial (n=9) anddistal colon (n=8) were simill.ar, aind themucosal addition of' amiloride decreasedshort circuit current (Isc) by 147+32RA/cm' (p<()-t()5) and totatl conductani(ce(G,) by 1(6±0(6 mS/cm' (p<().()5) in distalcolon, but only decreased Isc by 47+13jAA/cm' (p<)-01) aind G, by 1-2+0(4mS/cm' (p<()-05) in proxinial coloii. Incholine Cl Ringer, Isc decreased allmost tozero in both colonic segments, suggestingthat electrogenic Nia' transport was presentequ,ally in the proximal aind distal colon.Kinetic studies on the basolateral Na '-K'pump revealed that the maximum pumpactivity, pump affinity lor Nia', and thenumber of Nat' ions binding to each pumpsite, were similar in the proximal and distalsegments. T'hus, short circuit currentthroughout humain colon is Nat'-dcpeindenitand probably reflects in clectrogenic Naitransport proecss which is aimiloride-sensitivc only in the distill segmenit. 'I'hissegmcntatl vairiability in aimiloride sensi-tivity miay reflcct a segmental differeiicc inmucosal responsivencss to aldosterone.

Differential effects of pure mineralocorti-coid and glucocorticoid hormones on colonicNa+ and K+ transport

G I SANI)iDI AND F MC(i ONI (DeparIm'ni 0oMedicitne, Hope Hospital (OLi(iei of

A I1335


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Manchester School of Medicine), Salford)The glucocorticoid dexamethasone binds toglucocorticoid and mineralocorticoid recep-

tors in mammalian colon, and stimulateselectrogenic Nat transport, K' secretionand mucosal Na',K'-ATPase activity, but itis unclear which type of receptor is more

important in mediating these changes. Wetherefore used microelectrodes in vitro tocompare the effects of hyperaldosteronism(secondary to dietary Na' depletion) and a

pure glucocorticoid (RU 28362, Roussel;600(tg/10() g BW/day for three days) whichhas no affinity for mineralocorticoid recep-

tors, on Na- and K' transport in rat distalcolon. Compared with control tissues,aldosterone increased transepithelial volt-age five-fold (p<0)0)01), total conductance2-3-fold (p<0(00 1), hyperpolarised thebasolateral membrane by 11 mV (p<0(025),and decreased the apical/basolateral mem-

brane resistance ratio by 64% (p<0.0l). Incontrast, RU 28362 increased transepithelialvoltagc three-fold (p<0(001), had no effecton total conductance or basolateral mem-

brane voltage, and decreased apical/basolateral membrane resistance ratio by60%X (p<0.01). Studies with specific ionchannel blockers indicated that both aldo-sterone and RU 28362 induced appreciableNa' and K' conductances in the apicalmembrane. Kinetic studies of the baso-lateral Na'-K' pump revealed that aldo-sterone increased maximum pump activityby 230% (p<000l), but RU 28362decreased maximum pump activity by 45%(p<O0 1). Thus, aldosterone and RU 28362both enhance apical membrane conduct-ance to Na+ and K+, but have oppositeeffects on the basolateral Na+-K' pump.

The transport effects of dexamethasone are

therefore likely to mainly reflect activationof mineralocorticoid receptors.

Mechanism of laxative action of phenol-phthalein of intestinal fluid absorption

ANDREA J PHILLIPS AND HELEN I LEAT[IAR[)

(Department of Pharmacology, CharingCross and Westminster Medical School,London) The action of phenolphthalein on

intestinal fluid transport was invcstigated,on rat everted intestinal sacs, by addingdrugs to the fluid bathing the mucosalsurface. Phenolphthalein (3-1 x 1)' M)reduced net fluid absorption in the smallintestine (p<0-05, n=25), but not in thecolon (p>0-05, n =7). The former effect wasantagonised by tetrodotoxin (3-1 x t) M,p<005, n= 10), indicating the involvementof a nervous mechanism. Atropine (3-5x

I( " M) partially antagonised phenol-phthalein whilst abolishing basal secretion(p<0.05, n=9), showing that cholinergicand non-cholinergic secretory nerves may

be involved. In the presence of atropine.imitation of the cffects of phenolphthaleinby carbachol (I- xIO10 M) suggests thatactivation of neuronal cholinergic receptorspromotes the release of a non-cholinergicsecretory neurotransmitter. Apamin(5x " M), which had no effect on basalsecretory activity (p>)-05, n=9), com-

pletely inhibited phenolphthalein (p<0(05,n=8). As apamin has been reported toinhibit neural release of vasoactive intest-inal polypeptide (VIP), this suggests a rolefor VIP as the non-cholinergic secretoryneurotransmitter. It is concluded that thelaxative action of phenolphthalein is parti-ally mediated by inhibition of small intest-inal fluid absorption via a nervous pathway.This appears to involve cholinergic activa-tion of a VIP-ergic nerve.

Lung permeability in inflammatory boweldisease (IBD) and coeliac disease (CD)

D A F ROBERTSON, H SIDHU, N TAYLOR,

A BRHTTEN, C L SMITH, AND G HOLDSTOCK

(Depts of Medicine anid Nuclear Medicine,Southampton General Hospital, Souithamp-ton) Respiratory disease and subclinicalpulmonary abnormalities have been identi-fied as complications of both CD and IBD,but the pathogenesis of the lung diseaseremains uncertain. Possible explanationsinclude a common mucosal permeabilitydefect allowing absorption of antigenswith subsequent damaging tissue reactions,both in gut and lung, or exposure to an

environmental agent causing both lesions,

immune complex deposition or pulmonaryvasculitis.We have studied lung function including

permeability measured by clearance ofinhaled 99TcDTPA (expressed as time forl/2 inhaled 99TcDTPA to clear from lungs toblood: t1,) in 25 IBD and 18 CD patients ona gluten free diet, and in 20 controls, allwithout respiratory symptoms. In IBDthere was evidence of obstruction to airflow(mean FEV1/FVC=75-80/O, control 810%p<0)-05) but no change in pulmonarypermeability (ti,= 70(3 mins v 69-2).

In CD airflow was not significantly differ-ent from control (FEV1/FVC=80%) butthere was an increase in pulmonary perme-

ability (tl,=48.9 min; p<0)(1). These find-ings suggest that the mechanism of lungdisease in CD differs from that in IBD, andsupports the hypothesis of a common

mucosal defect in lung and jejunum in CDallowing increased permeability.

Mucosal cell receptors and enteric nerves

mediate the spasmogenic effects of bacterialN-formyl oligopeptides on guinea pig ileum

V S CHADWICK, C H IJOBSON, ANt) M F BROOM

(Wellcome Medical Research Institute,University of Otago Medical School, POBox 913, Dunedini, New Zealand) Themechanism of the intestinal spasmogenicaction of bacterial N-formyl (methionyl)oligopeptides has not been defined. We havedetermined pA2 values for eight potentialantagonists (receptor, nerve or ganglionblockers) of N-f-met-leu-phe (FMLP) usinghistamine, acetyl choline, nicotine, 5HTand Substance P as control agonists. Asynthetic FITC-labelled peptide and a

rhodamine labelled anti-FITC antibodywere used to identify receptors for bacterialpeptides in ileal mucosa.

Atropine (pA2=8-4), pirenzepine (8X0)and tetrodotoxin (7-9) were potent antago-nists of FMLP induced contraction suggest-ing involvement of Ml cholinergic neuralpathways which modulate slow excitatorypost-synaptic potentials in the entericnervous system. Tachyphyllaxis experi-ments demonstrated that 5HT and Sub-stance P receptors were not involved.Salazopyrine, known to block FMLPreceptors on neutrophil leucocytes alsoantagonised FMLP responses (pA2=7-3).FITC-labelled peptide was bound and inter-nalised by granulated cells in ileal mucosa

and removal of mucosa from ileal segmentsabolished muscle responses to FMLP butnot to histamine or acetyl choline.We conclude that bacterial N-formyl

peptides, produced by many intestinalbacteria bind to cells bearing receptors in

ileal mucosa and produce muscle contrac-tion via enteric neural pathways.

GASI RODUODENAI POSTIRS

Enprostil - an advance in gastric ulcertherapy?

A G MORGAN, C PACSOO, P TAYLOR, AND W A F

MCADAM (Endoscopy Unit, Airedale GeneralHospital, Steeton, Keighiley, Yorks) In arandom double blind endoscopically con-trolled study, 100 patients with gastriculcers were treated with either enprostil (7t)

A1336


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The BritisIh Societv of Gastroenterologyv

p.g twice daily) or ranitidine (150 mg twicedaily). Demograiphic data wals evenlymatched. The healing rates at four, eight.and 12 weeks were enprostil 60%O. 90%. and940/ and for ranitidine 65%o, 90%o. and1000(0o rcspectively. Diarrhoea was acommon side effect of enprostil therapy(25%).To assess the effectiveness of mainten-

ance enprostil therapy, half of the patientswere followed for one year without treat-ment, and the others were given enprostil 70)[tg nightly. Endoscopy was done in bothgroups atfter six and 12 months, or if ulcersymptoms rcturned. The recurrence rate atsix and 12 months following ranitidine was670O and 750O and after enprostil 5t)00 and61%/ respectively. On maintenance enpro-stil, the recurrence rate was reduced to 30%oaind 38%, respectively. Seven of 48 patientsstopped treatment because of diarrhoeaior abdominal cramps. So far, eight of 12patients have relapsed within six months ofstopping maintenance treatment.

Enprostil is as effective as ranitidine inthe healing of gastric ulceration, but at the7t) Fg dosage, diarrhoea is a problem.

A double blind randomised multicentrestudy comparing Maalox TC tablets andranitidine in the healing of duodenal ulcer

J 0 IHUNTER, J (ROWE, R R GIl IF.S, K R GOUGIH,S LORBER, ANI) R J WAILKER (Addenbrooke'sHospital, Cambridge, Mater MisericordiaHospital, Dublin, Ottawa Civic Hospital,Canaida, Roval Uniited Hospital, Bath,Temple University Medical Centre, Phila-delphia, USA, anid Waltoni Hospital, Liver-pool) Patients with endoscopically confirmedsymptomatic duodenal ulcer were eligiblefor a double blind randomised multicentretrial comparing the antacid, Maalox TC(three tablets one hour after meals and atbedtime) with ranitidine (15t) mg twicedaily). The randomisation was stratifiedaccording to smoking history. The sympto-matic effects of treatment were recordedafter one, two, and four weeks. Repeatendoscopy was done after four weeks when,if the ulcer was unhealed, treatment wascontinued until a final endoscopy two weekslater.

Seventy nine patients were admitted tothe trial, 43 receiving Maalox and 36 raniti-dine. No significant demographic differ-ences were detectable between the twogroups. Thirty five patients taking MaaloxTC and 28 ranitidine completed the trial,the respective healing rates being similar atboth four weeks (67.7%) and 73.30%) and at

six weeks (829"% and 89X3(%). Patients whosmoked were less likely to heal (p<0)-00 1).No significant difference wias detected in theseverity ot ulcer symptoms between the twogroups. Eight patients defaulted and eightwere withdrawn because of adverse drugef fects, predominantly diarrhoea (fiveMaalox 'I'C, one ranitidine). Mlaalox TC is aconvenient antadcid formulation which iscomparable in efficacy to ranitidine in thehealing of duodenal ulcer.

Esaprazole increases gastric alkalinesecretion in man

M (USiLANDI (Institlito di Medicinia Ititertia,Milatno, Ital') Esaprazolc, N-I(N-cyclohexyl-carbamoyl)methylI piperazinehydrochloride, is a novel antiulcer com-pound reportedly as effective as cimetidinein the short term treatment of peptic ulcer.The drug, which has a low antisecretoryactivity, was found to exert cytoprotectiveef fects against various necrotising agents. Inorder to further elucidate the mechanism ofaction of esaprazole, the effect of the drugon gastric bicarbonate secretion has beeninvestigated in man. Twelve healthy volun-teers, 1i) men and two women, aged 18-56years, entered the study after givinginformed consent. They were treated withesaprazole by oral route at a dose of 1350mg daily for 10 days.

Before and after treatment basal bicarb-onate secretion in fasting gastric juicewas determined by means of Feldman'smethod, on the basis of H' concentration,secretory volumes, osmolality of gastricjuice and plasma. A highly significantincrease (p<)-()l) in gastric HCO3 secre-tion, from 2-65±0-57 (mean±SD) to4-28±0-79 mmol/h was observed afteresaprazole treatment, this effect beingdetectable in all subjects. -

Gastric HCO, ions are known to con-tribute to mucosal protection by neutralis-ing back-diffusing H' ions within the mucuslayer. Strengthening of the gastric mucusbicarbonate barrier by esaprazole maypartially account for the cytoprotective andulcer-healing properties of this new thera-peutic agent.

Colloidal bismuth subcitrate (DE-NOL) innon-ulcer dyspepsia. Placebo controlledtrial with particular reference to the role ofCampylobacter pyloridis

T ROKKAS, C PURSEY, N A SIMMONS. M I FIIIPE,ANI) G 1. SILAI)EN (Gastroenterology Unit, Div

of Medicine, Dept of (liCnical Bactcriologi'and1{ Histopathologv%, UMDS Gu(,'s and SIThoias' Hospitals, l ond(lonl) Fifty patients(35 M, 15 F. meaitt age 39S5, raintge 20-60yeairs) with NUIt) aind no recenit history ofNSAID contisumptioni. were studied toevaluate the effect of De-Nol on svymptomiisand on infection with C p) 'lori(li.. Allpattients had normal abdominal U/S and inall gastroscopy revealed t10 significantG1 pathology. Patients were allocatedrandomly to eight wceks treatment witheither placebo or De-Nol after which theywere re-evaluated clinically and enido-scopically. Antral biopsies were examinedbacteriologicailly and histologically.

Patients on platcebo aindi c-Nol did notdiffer regarding sex, ige. smoking, alcoholconsumption and C pvloridiis +ve cultures(1 1/25 in De-Nol and 8/25 in plaicchogroup). All 19 C(P!loridiis +ve patients100%') had active gastritis in contrast with

only cight ( 193(%) of C p)ilorI(is -vepatients (p<0)001 ). Dc-Nol eradicated Cpyloridi. from 9 (81X .8o) Of the 11 +vepatients compatred with placcho which didnot eradicate C pvyloridi.s from any of theeight +ve patients (p<0-01). In patientsreceiving De-Nol, compared with placcho.the gastritis (p<0-01 ) and the symptoms(p<t)-001) improved.We conclude that De-Nol is eflective in

NUD aind this is associated with eradicationof CpYloridiis from the stomalch.

Eicosanoids in gastroduodenal mucosa ofduodenal ulcer (DU) patients treated withcolloidal bismuth subcitrate (CBS)

A AIIMED1), S K CAIRNS, 1) VAIRA, ANI) P KSAL MON (Dept of (as.itroenterology', Middle-sex Hospital, London) The role of cytopro-tective prostaglandins (PG) in the patho-genesis of DU remains controversial, butleucotrienes ( LT) probatbly play a pro-inflammatory role. We hatve measured theformation of immunoreactive PGE2 aindLTC4 in mucosal biopsics taken from 13patients with DU, both before and fourweeks after treatment with CBS. In allcxcept one, ulccr healed aftcr therapy.Multiple duodenal and antral biopsieswere incubated at 37°C in prewarmed/oxygcnaited (95%, 02+5/% CO,) Tyrodesolution (basial release) and trainsferred to atsecond incubation contatining ionophonc5 !tg/ml A23187 (stimulated release).Radioimmunoassay analysis of the super-

natant showed basal synthesis of PGE2 wassignificantly higher in antral thian duodenal

A1337


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Societv of GastroenterologvJ

mucosa (pg/mg WW/20 min, means±SEM)1198±157-4 and 887+102-3 (p<0-02),whereas release of LTC4 was significantlyhigher in duodenal than antral mucosa,96± 18-1 and 56+12-2 (p<0-()1). lonophorestimulated LTC4 generation was signific-antly increased (duodenum 183 ±44-3,p<0)1; antrum - 89+20-2, p<0-05), butnot the release of PGE, (duodenum -

278±111-7 (p<0.01l); antrum 329±86-8(p<0(OI) suggested a different cellularorigin for the two eicosanoids. Ulcer heal-ing after CBS therapy did not alter basalPGE2 and LTC4 generation either in duo-denal or antral mucosa, but stimulatedLTC4 formation was significantly reduced(p<0-OOI).These results suggest that LT formation

may be of greater significance than PGsynthesis in duodenal ulcer disease.

A single blind comparative study of miso-prostol to sucralfate and placebo in theprevention of aspirin induced ulceration

F LANZA, K F" PEACE, 1. GUSTITUS, AND

B DICKSON (Baylor College of Medicine,Houiston, Tv and C D Searle Skokie, II,USA) We report for the first time, theresults of a direct comparison of the cyto-protective properties of misoprostol (M) tosucralfate (S) and to placebo (P) in normalsubjects taking aspirin (A).

Thirty healthy volunteers with endo-scopically normal gastroduodenal mucosaewere randomised into three equal groups toreceive M 200 [tg, S 1 g, or P, co-administered with A 650 mg four times dailyfor seven days. After a single final dose onday seven, endoscopies were carried outand the mucosac were graded in a blindedfashion according to a 0-4+ scale: 0=normal; I + =single hemorrhage or erosion;2+=2 to 1t) haemorrhages or erosions;3+ = 11 to 25 haemorrhages or erosions; and4+ =more than 25 haemorrhages orerosions or an invasive ulcer of any size.

Utilising a score of 2+ or less as aclinically significant prophylactic success,the success rates were: M 10/10 (100)%),S 2/10 (20%)), and P 0/10 (0(%). Misoprostolwas statistically superior to S (p=-()OI()and to P (p=O-OOO(l). Ninety five per centconfidence intervals on the difference insuccess rates between M and S and M and Pwere (44%; 1O)/0O) and (61%, 100%)respectively. Adverse experiences wereminor and compairable across treatmentgroups.

Sucralfate (SUC) v cimetidine (CIM) for thetreatment of duodenal ulcer (DU) associatedantral gastritis

W M iUI, S K lAM, J 110, C 1. IAI, A S F LOK, M M T

NG, AND I.UI (DepartmenGits of Medicin]eand Pathology, University of Hong Kong,Queeni Mary Hospital, Honzg Kong) Antralgastritis is closely associated with DU andimproves with DU healing and treatmentwith misoprostol. We conducted a singleblind, randomised study to comparc thetherapeutic effect of CIM, an acid inhibi-tory agent (200 mg tds and 400 mg nocte)and SUC, aI site protectivc agent (1 g qid) onantral gastritis associated with active DU,two atntral and two fundal gastric biopsicswere taken during endoscopy beforc andafter treatment. The activity of the gastritisas assessed histologically by the infiltrationof polymorps was graded blindly by twopathologists as nil, mild, moderate orsevere. The two groups (CIM n=71, SUCn=69) were comparable in their clinicilIcharacteristics. Despite similar rattes of DUhealing in the two groups (75(YO, 78% at fourweeks) the incidence of improvement ofactivity of antral gastritis (nil or mild asendpont) was significantly higher in theSUC (33-3%) than in the CIM (183%/)group (p<0-05) and in the healed DU groupthe improvement was significantly higher inthe SUC (36-40/) than CIM (16-4%) group(p<t).02). Smoking did not affcct theimprovement in both groups. During thesubsequent 24 month follow up ulcer recur-rence ratte in the CIM group was double thatin the SUC group.We concludc that histological improve-

ment of the activity of DU associated antralgastritis was significantly higher with sucral-fate than with cimetidine treatment despitesimilar healing rate.

Does the pH influence the protective actionof sucralfate against mucosal injury?

B J / I)ANESII, A DUNCAN, AND R I RUSSEIT I

(Gastroenterology Unit, The RoyalInfirmary, Glasgow) An acid medium isclaimed to be mandatory for the mucosalprotective action of sucralfiate. In situationswhere atcid secretion is inhibited the efficacyof this drug to protcct against mucosalinjury has, therefore, been questioned. Weexamined the effect of sucralfate onmucosal erosions induced in rats (n=200)by aspirin (PKa 3-5) and bile acids (tauro-deoxycholic and taurocholic acids: PKa 1 8,and glycocholic acid: PKa 3.9) at pHs 1-5and 6-5. We also studied the effect of

sucralfate on the intragastric pH (using a pHmicroelectrode), on absorption of salicylateand on peptic activity. These parameterswere examined four hours after oral admin-istraution of test solutions. Sucralfate signifi-cantly reduced the incidencc and severity ofmucosal erosions (p<0-00l) induced byaspirin alone and when combined with thethrce bile acids at both pHs of 1-5 and 6e5.Rats which received sucrailfate had a higher(p<0)-00l) intragastric pH than those whichdid not receive this drug. Sucralfate did notinfluence salicylate aibsorption or pepticactivity. Wc have shown that the protectiveeffcct of sucralfate against mucosal injuryinduced by aspirin and bile acids is not pHdepcndent. The simultaneous use of drugswhich inhibit acid secrction, thus, may notreducc the protective efficacy of this drug.Stimulation of alkaline production may playa role but inhibition of pepsin and binding ofaspirin to sucralfate aire unlikely to bcfactors in the mechanism of sucralfatesaction.

Effect of treatment with famotidine onT-lymphocyte populations and activity induodenal ulcer patients

R MOUNTFORI). A W PRI ECF'I, R JONES, E GROVI,ANI) J C(OJTREI I (Universitv Department of'Medicine antd Oncologs Researcl Utit,Bristol Roval Infirmary, MSD Limited,Hoddesdon, Herts) Cimetidine has beenreported to stimulate cell mediatedimmunity, possibly through interaction withsuppressor cell H2-receptors; such effectscould contribute to ulcer hecaling with H.-antagonists. We have studied the effects otfamotidine, an H.-antagonist structuriallydifferent from cimetidine, on T-cell popula-tions in duodenal ulcer piatients. Lympho-cytes were isolated from blood samplesfrom 20 patients, before and during treat-ment with famotidine 40 mg nocte, andfrom untreated control groups of non-ulcerpiatients (n= 19) and healthy volunteers(n =6). Mean pretreiatment values for T-cell(pan-T) percentages and helper/suppressorratios, estimated using monoclonal anti-body labelling, did not vary significantlyfrom control groups (64±12% and 1-97+0-79:1 v 55±13% and 20)4±0+99:1 non-ulcer controls, 56±l13'o and 1-5±0+76:1healthy controls) or during treatment: threedays- 63+ 13%o and 2-0+0-75:1; two weeks- 64±1l1% and 1-X2+±0-X:; six weeks -

58+9°/,o and 1-84±0-53:1. Mean pretreat-ment Con-A stimulated lymphokine pro-duction, estimated by reduction of macro-phage electrophoretic mobility, was also

Al1338


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


sinilar to control values (13±+4(Xo non-

ulcer controls 12±50/). healthy controls8+3% ) and did not vary significantly duringtreatment: three days - 13±6Oo; two weeks- 13+6°/,, six weeks I l±5%. In conclusionfamotidine does not modify T-lymphocytepopulations or lymphokine production induodenal ulcer patients. Immunomodula-tion by cimetidine may be due to molecularrather than H-medilted effects.

24 hour intragastric acidity and plasmagastrin concentration in healthy volunteerstaking famotidine 40 mg nocte

S liANZON-Mit LER, R E, P'OUNDER, S G BALI., D J

DAtG(iLrEISH, J COWARD, AND A 0 JACKSON

(Royal Free Hospital School of Medicine,London) Famotidine is a new H2-receptorantagonist. This study compared the effectson 24 h intragastric acidity and plasmagastrin concentration of the 7th day oftreatment with either famotidine 40 mg or

placebo, taken as a tablet at 2100 h. Thetreatment was given in a predeterminedrandom order. Ten healthy volunteers tookpart in the study: median age 22 years,median weight 72 kg.Median integrated 24 h intragastric

acidity fell from 066 pmol/h/l on placebo to281 pmol/h/l on famotidine 40 mg nocte.Conversely, median integrated 24 h plasmagastrin concentration rose from 248 to 394pmol/h/l during treatment with famotidine.Both of these changes are significant(P<(d)0)).

Analysis of the data by 'meal-relatedintervals' shows that the decrease of intra-gastric acidity during treatment with

famotidine is due entirely to a decrease ofnocturnal acidity - daytime acidity is

unaffected by the drug. Famotidine 40 mgnocte causes a significant increase of plasmagastrin concentration during the night,morning and afternoon, with no effect in theevening.

Famotidine 40) mg nocte causes a signifi-cant decrease of nocturnal acidity, with a

longer lasting rise of plasma gastrin concen-

tration.

Optimising the effect of nizatidine on gastricpH: early evening meal together with thedrug is best

P DUROUX, P BAUERFEINI). C LMD)i, H R KOil/,I) MARGAILITII, G, DORIA, A KARPF-, P P

KEO0iANIE, ANt) A 1. BlUM (GastroenterologyCHUV-Lausanntie anid Stadt.spital Triemli-

Zluricl, Switzerland, Lilly--England) In our

previous studies H2-antagonists were more

effective after early evening intake thanafter late intake. The question arises

whether the meal interacts with the drug.We thus conducted a randomised doubleblind placebo controlled crossover study. Ineach of 12 healthy volunteers six ambula-tory 24-h pH-metries (glass electrode in thegastric corpus, 30) samples/minute, start at4 pm) were performed. Drug serum concen-

trations were determined in the first fourhours. The following six treatments were

randomised: N6/D6 (nizatidine 300 mg at6 pm and dinner at 6 pm), N9/D6 (nizatidine30)0 mg at 9 pm and dinner at 6 pm), N6/D9,

N9/D9, P/D6 (placebo, dinner at 6 pm) andP/D9, respectively. pH-medians were calcu-lated from 6 pm to 12 pm (evening), from6 pm to 7 am (night) and from 7am to 12 am(morning).

In the evening period pH-values were

higher with N6 than with P (mean values ofpH-medians: N6/D6 3-5 v P/D6 1-4; N6/D93-9 v P/D9 1-6); N9 was not different fromP; N9/D6 was inferior to NO (N9/D6 1-7 v

N6/D9 3-9 and N6/D6 3-5). During the nightN6 and N9 were superior to P (NO/DO 3-8and N9/D6 3-6 v P/DO 1-9; N6/D9 3-3 andN9/D9 3-5 v P/D9 1-5) (p<0-(0 multiplecomparisons by Wilcoxon/Wilcox). Theeffect of nizatidine ceased with breakfast.Drug serum concentration peaked l120)minutes earlier when nizatidine was takenwithout meal than when nizatidine and mealwere taken together.We conclude that (1) A concomitant meal

slows the absorption of nizatidine and may

thus prolong its action. (2) Nizatidine ismost effective when given at 6 pm togetherwith a meal. Supported by SNF 3.827.0.80and DFG Em 36/1-3.

Dose dependent inhibition of twenty fourhour intragastric acidity by WY-45,727 a

new potent H2-receptor antagonist

H MERKI, 1. WITZEI., J NEWMANN, D KAUF-

MANN, AND J ROEHMEL (INIRODUCLD BY R P

WAIT) (DRK Hospital, West Berlin, FRG)WY-45,727 (Wyeth Labs) is a new furanH2-receptor antagonist of greater potencythan ranitidine (3-lOX). We measured theeffects of increasing doses of WY-45,727 (20mg, 5() mg, 150 mg) on 24 hour intragastricacidity in 20 normal volunteers in randomorder under double blind conditions. pHwas measured continuously by intragastricglass electrodes and each volunteer under-went four identical studies. Medication was

taken after supper ait 183(0 h. Statisticalanalysis to assess the hypothesis thatincreasing doses would produce increatsilngsuppression of' acidity was hy the non-parametric PAGE test for order elects.Median nocturinlul pH rose f'roni 13(placebo) to 2-4 (92% decreiase ol icidity).

-(0 (99"/ decrease), aInd 6-7 (99% decreatse)with increasing doses of WY-45.727. Anti-secretory effects during the day werclimited; 21 'o, 37%, and 75%0 decreasesrespectively. Overall 24 hour median p11rose from 1-3 (plaicebo) to 1-9, 3 1, and 4l5with 20 mg. 40 mg. aind 150 mg respectivelv.Cleair dose dependenit decreasing aciditywas confirmed (p<0-0001) and eflects weregreatest at night. Such potent suppressionlof gastric acidity should make WY-45,727an effective ulcer therapeutic agent wortlivof further study.

Effect of intravenous BMY-25368-01 (BMY)and ranitidine (R) on meal stimulated gastricacid secretion in man

It G DAMMANN, F BURKiiARD)i, P MUiIiF-R, B

SIMON, R R (CR.NSIIAW, K MANGiHANI, ANt) S 1

SCIiWARJ/. (Kranikenthai.s Bellathiaen, Halm-blurg, FRG, a1/1(1 Pharmaceutical Researchand Dev'elopmnent Division, Bristol-Mver.sC(o, Wallingf.ord, USA) In this blinded(BMY doses v placeho), randomised, cross-over study, a singie iv dose of BMY (1-amino-2 [3-(3-piperidinomethylphenoxy)propylaminol cyclobut- -cne-3,4-dionchydrochloride), an inhercntly long actingand novel histamine H,-receptor antago-nist, 50 mg of R, or placcbo was adminis-tered to healthy voluntecrs. BMY wasstudied at 5, 12-5, and 25 mg. Subjectsreceived either BMY and placebo or R andplacebo. Six subjects were in each groupexcept the BMY-5 mg group that consistedof five subjects. Medications were adminis-tered at (07(00 hours, and uniform homo-genised meals (standardised to pH 5) wcrcadministered at 0800(, 13 00, and 180(t)hours. Volunteers receiving 25 mg of BMYwere administered a 4th meal at tIS t)0 hoursof the subsequent day. Gastric acid secre-tion was determined using an automatictitrator that maintained the intragastric pHat 5 for two hours after administration ofeach meal. No untoward clinical effectswere observed with either BMY or R.

Intravenous BMY is more potent thanand longer acting than R. A single iv dosc of25 mg of BMY produces significant inhibi-tion of meal stimulated gastric acid secre-tion for 24 hours. Intravenous BMY is a ODantisecretory agent.

A 13'39


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroeniterology

Omeprazole improves antral gastritis associ-ated with duodenal ulcer (DU)

W M [HUL, S K LAM, J 110, I lUI, W Y LAU, F J

BRANICKI, C I. [AI, A S F lOK, AND M M T NG(Departments of Medicine, Surgery andPathology, University of Hong Kong, andGovernment of Surgical Unit, Queen MaryHospital, Hong Kong) Antral gastritisoccurs in close to 100%(" of patients withactive DU. Whether gastric acid secretionplays a pathogenetic role in antral gastritis isunknown. We carried out a double blindrandomised trial comparing omeprazole(OME) 10 mg OM, 20 mg OM and raniti-dine (RAN) 150 mg bd in 270 patients withactive DU. Healing of the ulcer wasassessed weekly by endoscopy with at leasttwo antral biopsies taken to assess theactivity and degree of chronic inflammationhistologically by the degree of polymorphand mononuclear cell infiltration respect-ively and graded blind by two pathologistsas nil, mild moderate and severe. The sex,age and maximal acid output were compar-able in the three groups. DU healing at twoweeks was 77%, 86%, and 63% respect-ively, while that at four weeks was 95%,960%, and 93% respectively. The percent-age of patients with improvement in theactivity of gastritis in the four consecutiveweeks were 9%, 40%, 51%, and 530% forOME 1( mg (n=78), 14%, 42%, 49%, and53% for OME 20 mg (n=81) and 2%, 23%,30%, and 33% for RAN (n=82), and lifetable analysis: OME 10 mg v RAN andOME 20 mg v RAN, p<0(0I. The degreeof chronic inflammation showed similarchanges. Gastric acid and smoking did notaffect the improvement of gastritis. Thusdespite similar DU healing rates at fourweeks, OME 10 or 20 mg daily is signific-antly better than ranitidine 150 mg bd inimproving antral gastritis, suggesting thatintense acid inhibition helps to improve theantral gastritis.

Omeprazole (OME) v ranitidine (RAN) forduodenal ulcer - weekly endoscopic assess-ment

W M HUI, S K LAM, W Y LAU, F J BRANICKI, C L

lAI, A S F LOK, M M T NG, K P POON, AND PJ FOK(Departments of Medicine and Surgery,University of Hong Kong, and GovernmentSurgical Unit, Queen Mary Hospital, KongKong) To investigate the DU healingefficacy of two to four weeks' treatment ofOME, 270 Chinese patients with endo-scopically active DU were randomised to adouble blind controlled trial of OME 10 mg

daily (low dose), 20 mg daily (standarddose) and RAN 150 mg bd. Forty sixpotential factors affecting healing includingclinical and endoscopic characteristics wereprospectively obtained and healing wasassessed by endoscopy at weekly intervalsfor at least two weeks and up to four weeksif DU remained unhealed. The cumulativehealing rates per protocol analysis in theconsecutive four weeks were 43%, 77o,94%, and 95% for OME 1) mg (n=83),49%, 86%, 93%, and 96% for OME 20 mg(n=87) and 29%, 63%, 83%, and 93% forRAN (n 84) respectively. (Life tableanalysis OME 10mg v RAN, p<0(03, OME2(0 v RAN p<0(002.) There was a trendtowards more rapid symptoms relief withOME, particularly nocturnal pain duringthe first week. In the omeprazole treatedgroup (combined 10 and 20 mg) healing rateis lower in the smoker than non-smoker(p<0-0008), early than late onset (symp-toms before and after age 3(1 years respect-ively, p<0.02), remission >5 v <5 months(p<0(05) and high acid v normal (p<0(05)by life table analysis. We conclude OME 1()mg achieved similar healing rates as OME20 mg and both resulted in significantlybetter healing rates than RAN. Smoking,short remission period, early onset and highacid secretion adversely affected healing.

Omeprazole and ranitidine in the treatmentof benign gastric ulcer - an internationalmulticentre study

A WALAN, J P BADER, M ClASSEN, C B H W

LAMERS, D W PIPER, ANt) R RU FGERSSON

(Linkoping, Sweden; Creteil, France;Munich, FRG; Leiden, Netherlands;Sydney, Australia; M(ilndal, Sweden) Theaims of this international multicentre studywere to study the ulcer healing and toler-ability of four to eight weeks' treatmentwith omeprazole 2(1 mg once daily (om),omeprazople 4(1 mg om and ranitidine 15(1mg twice daily (bd) in patients with gastricand prepyloric ulcers. Forty five centres in13 countries contributed a total of 6(12patients of whom 2(13 received omeprazole2(0 mg, 194 omeprazole 4(0 mg, and 2(15raniti'dine.At four weeks, 59%O of patients in the

ranitidine group were healed comparedwith 69%o of the omeprazole 2(0 mg groupand 80% of the omeprazole 4(0 mg group. Ateight weeks, the corresponding figures were850%, 89%, and 960O, respectively. Thehealing rates were significantly higher withomeprazole 4(1 mg both at four and eightweeks compared with ranitidinc (p=)()()1)

and omeprazole 2(0 mg (p=0(12). Afteradjustment for an imbalaincc att entry interms of more patients with larger ulcers inthe omeprazolc 2(0 mg group, both omepra-zolc groups produced significantly higherhealing rates at four weeks compared withranitidine (omeprazole 2(1 mg v ranitidinep=)-()0(l, omeprazole 4(1 mg v ranitidinep<0(-()()(5). There wats a more pronouncedimprovement in ulcer symptoms during thefirst two weeks in the omeprazole groups. Asignificant difference was shown betweenomeprazole 4(1 mg and ranitidine (p=( ()2).All treatments were well tolerated andthere was no difference in the prevalence ofadverse events between the three groups.

In conclusion, omeprazole 2(0 mg and 4(1mg om have been shown to be superior toranitidine 15(1 mg bd in promoting healingof gastric and prepyloric ulcers, and therewas a better symptom relief with omepra-zole 4(1 mg.

Effect of omeprazole on postprandial gastricand duodenal pH in exocrine pancreaticinsufficiency

G MCIAUCHLAN, (i P ( REAN, AND K F L MCCOLI.

(Utniversity Department of Medicine,Western Infirmary' and GastrointestinalCentre, Southern General Hospital, Glas-gow) In exocrine pancreatic insufficiencyduodenal acidity (pH<4) may inactivateresidual endogenous enzymes and gastricacidity destroy oral pancreatic supple-ments. The effect of omeprazole on gastricand duodenal pH has been studied in sixpatients with steatorrhoea caused bychronic pancreatitis. Combined glasselectrodes (Radiometer GK28(1C) werepositioned in the second part duodenumand body of stomach and 24 h ambulatorypH measured after seven days on omepra-zole 4(0 mg/day, omeprazole 2(1 mg/day andplacebo given double blind in random orderwith two week washout periods. The drugswere taken orally at (1800( h and the twothree hours postprandial periods after themidday and evening meals were combinedand analysed.The percentage of the postprandial

period that gastric pH was below 4 rangedfrom 58%-1()()% on placebo; on omepra-zole 2(1 mg it was O%O in three patients and42%., 68%, and 86%/O in remainder and onomeprazole 4(0 mg it was (1% in five patientsand 76% in one. In the duodenum, thepercentage postprandial time that pH wasbelow 4 on placebo was 000/ in three patientsand 13%, 15%, and 26%Yo in others, onomeprazole 20 mg it was 0%/0 in five patients

A1340


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


and 5% in one patient and on omeprazole40 mg was (% in five patients and 14% inone patient. The two patients with thehighest gastric and duodenal acidities werethe most resistant to acid suppression withomeprazole.

In conclusion, some patients withexocrine pancreatic insufficiency appearrelatively resistant to acid suppression withoral omeprazole and even 40 mg mane maynot maintain postprandial duodenal pHabove 4.

Serum gastrin concentrations and gastricendocrine cell population during longtermtreatment with omeprazole in man

W CREUT7FELDT, R LAMBERTS, AND G

BRUNNER (Departmenti of Medicine,University of GCittingen, FRG, and Depart-ment of Medicine, Medical School,Hanover, FRG) Twenty four patients withchronic gastric or oesophageal pepticulceration including four with antrectomyresistant to 300 mg ranitidine daily and unfitfor operation were successfully treated withomeprazole (o) 20-60 mg daily for 13-21months. Fasting serum gastrin concentra-tions were monitored and multiple gastricmucosal biopsies carried out during gastro-scopy every three to four months and fixedin Bouins's fluid. Oxyntic mucosa wasstained with Grimelius silver technique forECL-cells, antral mucosa immunohisto-logically for G- and D-cells and analyzed bymorphometry.

Gastrin concentrations increased signific-antly during the first 12 weeks of treatmentfrom 103±+14 to 205±29 pg/ml; no furtherincrease was seen thereafter. Volumedensity of G- and D-cells in the antralmucosa did not significantly change. Thevolume density of argyrophil cells in theoxyntic mucosa increased from 0-49± 1% to0.97+0.-1"/O except in patients with antrec-tomy (0-23+0-04) but did not reach thevalues of patients with gastrinoma (1-3±0-23'., n=7). No clusters of argyrophil cellswere observed.

It is concluded that hyperplasia of oxynticargyrophil cells is mediated by the hyper-gastrinacmia in man and antral G- and D-cell volume density does not change duringlongterm omeprazole treatment.

Omeprazole 20 mg om and ranitidine 150mg bd in the healing of benign gastric ulcers- an Italian multicentre study

L BARBARA, A SAGGIORO, J OLSSON, M

CISI ERNINO, ANI) M FRANCEiSCHI (INIROD)UCEI)

BY L BARBARA) (Ninleteeni Italiaii gastro-eniterological departments) The aims of thistrial were to study the ulcer healing, relief ofsymptoms and the tolerability of four toeight weeks' treatment with omeprazole 2(0mg om and ranitidine 15(0 mg bd in patientswith benign gastric ulcers (GU).

Nineteen centres participated with a totalof 84 patients randomised to omeprazoletreatment and 83 patients to ranitidinetreatment. Four and three patients in eachtreatment group were found to have malig-nant ulcers and were consequently excludedfrom all analyses.The healing rate was significantly

(p<0(05) higher on omeprazole (35%74o, and 96%/O) than on ranitidine (9%,53%, and 85'S) after two, four and eightweeks, respectively. At entry a relativelylarger portion of the patients in the omepra-zole group had small ulcers (<10 mm).When adjustment WaS made for thisomeprazole was still found to be superior toranitidine in the healing of GU at all pointsof observation in the Intention to Treatanalysis. Symptoms were rapidly relievedon both treatments. Relief of symptoms wasfaster and more extensive in the omeprazolegroup compared with the ranitidine group.Both treatments were well tolerated andonly few adverse events were reported.Omeprazole 20 mg om has been shown

superior to ranitidine 150 mg bd in thehealing of benign gastric ulcers.

Omeprazole versus ranitidine in duodenalulcer patients unhealed after six weeks treat-ment with H2-receptor antagonists

FRENCHI COOPERATIVE STUDY (INTRODUCEDBY A COR I1OT) (CHRU de Lille, France)We studied 151 patients with endoscopic-ally proven duodenal ulcer (greater axis>5 mm) unhealed after at least sixweeks of treatment with either cimetidine800 mg (61 patients) or ranitidine 3(M)mg (90 patients) daily. The study was per-formed double blind and patients wererandomly allocated to either omeprazole 2(0mg om or ranitidine 150 mg bid. Endo-scopies were done on day 15 and (ifunhealed) on day 29. Endoscopic healingwas defined as total reepithelialisation of(all) the ulcer site(s). Before breaking thecode 26 (day 15) and 36 (day 29) patientswere excluded from the per protocolanalysis.

Healing rates after two and four week'stherapy were 48% and 80X% for omeprazoleand 46% and 75% for ranitidine (NS).Relief of ulcer related symptoms was rapid

in both groups. A lower healing rate wasobserved in patients with larger ulcers andlonger treatment with H, receptor antago-nist prestudy.

It is concluded that both omeprazole andranitidine treatment resulted in high healingrates after two and four weeks in patientsunhealed after six weeks treatment with H.receptor antagonist.

What really happens during maintenancetherapy of duodenal ulcer?

F J S BOYD, J G PENS'I'ON, I) A JOiiNSiONI, ANI)

K G WORMSiEY (Department of TllTherapeutics,Ninewells Hospital, Dltidee) Current main-tenance trial protocols do not evaluate theincidence of DU recurrence, the propor-tions of symptomatic to asymptomaticulcers at the time of recurrence, nor theproportions of asymptomatic ulcers whichsubsequently reheal, remain unchanged orcause pain.We have endoscopically re-examined 34

patients at monthly intervals for up to oncyear during maintenance treatment of DUwith ranitidine 15t) mg ait night. If a painlcssulcer was detected treaitment and lollow upremained unchanged. We stopped the studyin patients whose ulcers were associatedwith pain. The mediain follow up was cightmonths (range 1- 12).The incidence of recurrent ulceration was

34/, at six months and 46%/ at 12 months.The ratio of asymptomatic to symptomaticulcers at detection was 5:1. Averagemonthly probabilities for events of interestin asymptomatic ulcers were: rehealling33%; unchanged 61%'; and pain 6%. Thespontaneous rehealing rate of 33%, inasymptomatic ulcers means that in a 12month trial with endoscopy of asympto-matic patients at six and 12 months, recur-rence rates would be underestimated by47%, and 33% at six and 12 months respect-ively. In order to provide accurate informat-tion on the incidence of ulcer recurrence,current maintenancc trial protocols shouldbe modified.

Effects of 'mucosal protective' and anti-secretory drugs on assimilation of foodbound vitamin B12

J I HANSI.IP, D GIDDEN, F, J S BOYD), N MARKS,AND K G WORMSILFY (Departmenit of Thera-peutics, Ninewells Hospital, Dundee, andGastrointestinal Clinic, Groote SchuuiirHospital, Cape Town, S. Africa) Normalassimilation of food bound vitarmin B13X

A 1 34 1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


appears to require adequate acid and pepsinsecretion, in addition to intrinsic factorsecretion, by the stomach. Drugs whichinhibit gastric secretion, or have non-specific B12 binding activity (R-factor-likeactivity), or bind with B,2-intrinsic factorcomplex may thus impair assimilation offood-bound vitamin Bl,. We have studiedthe ability of a range of 'mucosal protective'agents to bind with B12 or B12-intrinsicfactor complex in vitro, and the effects of arange of 'mucosal protective' agents andgastric antisecretory drugs on the assimila-tion of egg yolk bound vitamin B,2 inpatients undergoing treatment for duodenalulcer disease.

De-Nol, sucralfate, carbenoxolone andantacid-alginate preparations bound eithervitamin B,2 or B,2-intrinsic factor complexin vitro, but these preparations did notreduce the assimilation of egg yolk boundvitamin B,2. Both ranitidine 15t) mg andcimetidine 400 mg administered 90 minbefore the test meal, or ranitidine 300 mgtaken on the evening before the test mealsignificantly reduced the assimilation ofegg-yolk bound vitamin B,2. Vitamin B,2malassimilation is a possible consequence oflongterm therapy with antisecretory drugs,but not with 'mucosal protective' agents.

Effect of bismuth subcitrate on gastricbicarbonate secretion in vitro and in man

C J SHIORROCK, J CRAMPION, L GIBBONS, AND

W D W REES (Hope Hospital, Salford) Themode of action of colloidal bismuth inhealing peptic ulcers remains unclear. Wehave examined the action of bismuth sub-citrate on gastric bicarbonate secretion; animportant component of mucosal defence.Addition of bismuth subcitrate (1() to 1t)M) to the luminal side isolated amphibianmucosa, mounted in a Ussing chamber,produced a dose dependant increase inbicarbonate secretion (1() hM: 0-19+±004 to0(23±0+05: I()- M: 0 17±0-4 to 0-24+0-04;It) M: 0-28+0-05 to 0-47+0-06 ,mol/cm2/h. All means +SE, n=6 and p<0-05).

Using a previously published perfusiontechnique the effect of bismuth subcitrate(1-10 mg/ml) on human gastric bicarbonatesecretion was studied. Bicarbonate secre-tion was calculated from the pH and pCo2 ofgastric aspirates while acid secretion wassuppressed by intravenous ranitidine. Aftera basal hour, bismuth subcitrate (1-10mg/ml) was added to the perfusate andbicarbonate secretion measured for afurther hour. At concentrations within thetherapeutic range, bismuth did not increase

bicarbonate secretion (mean+SE, n=6:I mg/ml=593± 118 to 611 + 163: 10 mg/ml=453±114 to 412±86 [tmol/h).These results suggest that stimulation of

bicarbonate secretion by bismuth does notoccur in man despite being demonstrable invitro.

F- NDOSC(OPY

Potential for endoscopic therapy in primarysclerosing cholangitis

S R CAIRNS AND P B cOFrON (Department ofGastroenterology, The Middlesex Hospital,London) Over 36 months, 17 of 23 con-secutive patients diagnosed PSC receivedendoscopic therapy. Treatment was sphinc-terotomy (SPfi) alone in five, nasobiliarydrainage (NBD) four (two with SPH), intra-biliary steroid infusion (I-BS) five, (threewith SPH), endoprosthesis (EP) 10 (fivewith SPH). Balloon dilatations were per-formed in four patients (three EP, oneI-BS).

Sphincterotomy alone produced a meanfall of bilirubin (Bil) 320O, alkaline phos-phatase (ALP) 29%, over 1- 11 months(mean six) follow-up. Nasobiliary drainageprovided clinical and biochemical improve-ment in all patients, follow up three to ninemonths (mean six) with mean fall in Bil39%, ALP 36%. I-BS was attempted in fivepatients, successful in four, with meandecrease in Bil 59%'O, ALP 32%, follow up1-28 months (mean 16). Endoprosthesisplacement succeeded in nine of 10 patientsand lowered Bil concentrations (meandecrease 4t)%) and ALP (mean decrease27%h) in eight patients with substantialsymptomatic improvement in six, follow up1-28 months (mean nine).

Six of nine patients given EP remainedwell without recurrence of cholangitis withstable LFF's up to one year after stentextraction.These results suggest that endoscopic

therapy may be useful, at least in the shortterm, providing biochemical and clinicalimprovement in most patients with sympto-matic PSC.

Randomised comparison of Nd YAG laser(L), heater probe (HP) and no endoscopictherapy (C) for bleeding peptic ulcer

K MA1-I1EWSON, C P SWAIN, M BLAND, J S

KIRKHAM, S G BOWN, ANI) r C NORTHFIEILD

(Gastroenterology Units, St James's atidUniversity College Hospitals, Lon,don,) TheNd YAG laser can significatntly reducerebleeding and mortality rates in bleedingpeptic ulcers, but less expensive electrodesystems might be equally effective. Thisstudy is a randomised comparison of thebest electrode (HP in our animal experi-ments) with the best laser (Nd YAG in ourclinical trials) and no endoscopic therapy.One hundred and forty three consecutivepatients with a bleeding peptic ulcer andstigmata of recent haemorrhage (SRHI)accessible to endoscopic therapy wereincluded. Stratification was according to thetype of SRH, and randomisation was bialsedtowards the new therapy (HP). The diagno-sis of rebleeding and subsequent manage-ment were by a clinician unawatre of thetherapy given. The rebleeding rate wassignificantly lower in L (20%) than C (42%O)(p<0-05), but in HP (28%) was not signific-antly different from either of the othergroups. There was no significant differencein mortality rate between the three groups(L=2%, C=9%, and HP= I%O). Coin1-bined analysis of these results with ourprevious randomised Nd YAG laser trialwith identical protocol but only two wayrandomisation showed significantly fewer Lthan HP rebleeds ( 14%° v 28%, p<0(05) anddeaths (2% v 10%. p<0(05). The resultsconfirm Nd YAG laser but not heater probeefficacy, and the combined results showsignificantly superior results with the NdYAG laser than with the heater probe.

Multipolar electrocoagulation (MPEC) inthe treatment of ulcers with non-bleedingvisible vessels (VV): a prospective, control-led trial

1. IAINE (INIROI)UCID 13Y .I VAIIN/UIILA)(USC School oj Medicine, Lose A igeles,CA, USA) This study assesses the efficacy olMPEC in reducing the morbidity ol ulcer'swith visiblc vessels (VV). Patients withupper G6 haemorrhage aind (a) blood trans-fusions of -2 units in 12 hours: (b) systolicBP s9() mm Hg, heart rate -11( beats/min, or orthostatic chainges in systolic BP-2() or heart rate :2(); or (c) hacmatocritdrop -6%M in 12 hours were considered lorentry (n=466). Patients werc ralndolisedto MPEC or sham MPEC if endoscopyshowed a VV in an ulcer without activebleeding (n=60). The patient and hisphysicians were blinded regiarding treat-ment.The treatment and control groups haid

no significant differences in any ot 21

A 1342


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The Britisi Society of Gastroenterology

characteristics evaluated at the time ofrandomisation. Results after entry for sham COLORECTAI-MPEC v MPEC: rebleeding: 12/29 (410%) v6/31 (19%o); blood transfused: 3-0±0+7 v Acute septi1.7±0.4 units; emergency surgery: 8/29 disease. A(28%) v 3/31 (10%); hospital days: 5-9+07 survvalv 4-3±0+4; and cost: $5270+680 v $3780+480. Although these results appear to A P CORDERfavour MPEC therapy. the differences are PROFESSORnot statistically significant. One death Uniet F Levoccurred in the MPEC group (without Ceneral Frebleeding); no patient died in the control Ipswich aHGgroup. hundred an

This interval report of an ongoing trial is nosis of. cencouraging with regard to the use of undergoingMPEC in ulcers with non-bleeding VV. A suspectedlarger sample size will be required, how- abdominalever, to establish statistical significance in retrospewtivfavour of MPEC. tonitis whic

one per cercolonic pernabscesses atsix patients

Histology and cytology compared for the exteriorisatidiagnosis of neoplastic lesions detected at proximal cocolonoscopy drainage or

inpatient doJ J r TATE, A hIERBERT, AND (G T ROYIE overthesett(INTRODUCED BY I IAYILOR) (University between th(Slurgical Unit and Departtnent of Pathology, ment and seSouthanipton General Hospital, Southamp- comparisonton) Increasing use of fibreoptic endoscopy computer n

in the investigation of the large bowel allows ance of tre,biopsies to be obtained more often than along with Ipreviously. The small size of endoscopic age, medicabiopsies, however, may be less than ideal antibiotic trfor histological diagnosis. We have com- age in survipared cytology of colonoscopic biopsy those treatespecimens with histological analysis. when other

Seventy two biopsy specimens were this way (pobtained at colonoscopy in 38 patients using trolled trialha standard 2 mm diameter biopsy forceps. A advantagecontact smear was made from the biopsy treatment cspecimen onto a microscope slide, this diverticularbeing fixed immediately. The same tissuesample was sent for routine histologicalexamination. National auOne sample submitted for histology. and disease

four submitted for cytology were unsuitablelor cvaluation. The final diagnosis was R G TUDORcarcinoma in 15 samples, benign adenoma RESEARCIIin 25 and normal mucosa in 32. The sensi- General Jtivity of cytology for carcinoma was 87%. Birminghan'I'he sensitivity of histology for malignancy ing 32 hospwas the same. Both techniques were 100% dom havespecific. When results of biopsy and document tcytology were combined the sensitivity complicateoincreased to 1t)00). Two hundr4

Cytology of colonscopic biopsies may be with a meacarried out simply and rapidly and is a useful women. Sixadjunct to conventional histology. The need previous hto repeat biopsies of suspicious lesions is One hundravoided, managed b)

iccomplications of diverticularradical approach improves

R, J WILLIAMS (INIRODUCED BY

TAYLOR) (University Surgical'el, Centre Block, Southamptonlospital, Southampton, andospital, Ipswich, Suffolk) Oned two patients with a final diag-mplicated diverticular diseaseurgent laparotomy because ofperitonitis (n=81) or intra-abscess (n=21) were studiedvely. Sixty two patients had pern--h was faecal in 16 cases. Fortynt of patients had macroscopicforations, 33% intra-abdominalrid 42%0 colonic masses. Thirtys were treated by excision orion of the affected colon, 41 bylostomy and drainage and 25 bylaparotomy alone. Although 21leaths were distributed evenlyhree groups, marked differencese groups in age, antibiotic treat-everity of pathology made directis unhelpful. Therefore, using amodel, the prognostic signific-atment group was investigated12 other features which includedal history, operative findings andreatment. A significant advant-ival to discharge was found fored by excision or exteriorisationfactors were controlled for in

)<0(05). In the absence of con-ls, this is the best evidence of anfor a radical approach in theDf acute septic complications ofrdisease.

dit of complicated diverticular

ON BEtIALF OF 1 HE SURGICALSOCIFTY 'COILON CLUB' (TheHospital, Steelhouse Lane,mn) Thirty six surgeons represent-)ital centres in the United King-collaborated to prospectivelythe morbidity and mortality ofd diverticular disease (CDD).ed patients have been recruitedn age of 67-8 years; 129 werety six per cent presented with noistory of diverticular disease.red and twelve patients werey surgery, and in 100 a resection

was performed. Primary anastomosiswas achieved in only 47"%, however.Laparotomy findings were often at variancewith the pre-operative assessment. Woundsepsis developed in 27.6%0 of cases, post-operative abscess in 15 patients, anasto-motic dehiscence in 12 8%Yo and stomal com-plications in 17.5%. Of the 105 pathologicalspecimens available, pathology in additionto CDD was recorded in 19 cases includinginflammatory bowel disease (four), andcarcinoma (one). There were 19 deaths inseries but only nine were caused primarilyby CDD. Faecal peritonitis carried thegravest prognosis (57% mortality rate). Theaudit has to date identified the necessity forprecise definition of the complication andpathology. The need for a larger study hasbeen highlighted in order that recom-mendations may be made about optimummanagement for each complication.

Is colectomy for severe idiopathic constipa-tion a success?

M A KAMM, P R IHAWILEY, AND J E LENNARD-

JONES (St Mark's Hospital, City Road,London) The experience of colectomycarried out for severe idiopathic constipa-tion in one hospital over the past 15 years ispresented.

Patients with Hirschprung's disease,megarectum or megacolon, and secondarycauses of constipation were excluded. All 44patients were women and the mean age atoperation was 34 years. The mean pre-operative stated bowel frequency was onceper 4-7 weeks, and abdominal pain (presentin 970Yo) and bloating were prominentfeatures. Eighty eight per cent used laxa-tives preoperatively. All patients had anormal barium enema. Colonic transit wasprolonged in 33/34 patients. Thirteen of 20(65%) had inappropriate puborectalis con-traction on EMG with straining, and 22/30(73%) were unable to expel a water filledballoon from the rectum in a test of simu-lated defecation.

Eleven patients had a caecorectal and 33patients had an ileorectal anastomosis.Mean follow up was 4-0 years (range twomonths to nine years). Postoperatively,two-thirds of patients were able to stopusing laxatives but 19 patients had diarr-hoea (>2/day) and this usually persisted.Pain persisted in 70(% of patients andremained a major problem. In a third bloat-ing and straining persisted. During followup five patients developed bowel obstruc-tion and in five constipation recurred.Eleven patients underwent further surgery

A 1 343


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


because of incapacitating diarrhoea or per-

sistent constipation.Preoperative physiology studies did not

correlate with outcome. Colectomy usuallyrelieves constipation in these patients butpain and other symptoms often persist.

Faecal diversion in the management ofCrohn's disease

M C WINSLEI, ti ANDREWS, J AlEXANDER-

WIl.LIAMS, R N AllAN, AND) M R B KFIGIII Y

(General Hospital, Birmingham) The roleof faecal diversion in the management ofCrohn's disease has been reviewed by retro-spective analysis of 85 patients who havehad 95 defunctioning procedures. Fifty onepatients had an elective resection with a

stoma to protect an anastomosis (n=38) or

for co-e .nt colitis (n=13). Forty fourpatients .. elective diversion alone as theprimary treatment for proctocolitis (n=23)or perianal disease (n=-l). Significantsustained remission was obtained in bothgroups (D=31 (70%) p<0(0l, R&D=42(82%Yo) p<0*0)01, median follow up 28 (6-24months). A significant number of patientsno longer required steroid therapy postdiversion (D= 14/16 p<0Ol1, R&D=21/30,p<0.01). There was a significant improve-

ment in post operative serum indices(D=ESR p<0-05) R&D=Hb p<0-)1,albumin and ESR p<0 05. Fifty eightpatients have undergone or await restora-tion of intestinal continuity, proctocolec-tomy has been carried out in 31 and threehave died. Disease relapse occurred in 16 ofthe 29 patients with continuity restored.Faecal diversion used to facilitate or limitresection, for steroid side effects, to avoidor delay proctocolectomy and for perianaldisease. The high rate of remission pro-duced by diversion suggests a faecal factorhas a role in the pathogenesis of Crohn'sdisease.

Abnormal vagal function in the irritablebowel syndrome

H L SMARI AND MICHAEI ATrKINSON (Depart-

ment of Surgery, University Hospital,Nottingham) Oesophageal symptoms inirritable bowel syndrome (IBS) arise frompathological gastro-oesophageal refluxoften associated with oesophagitis. Asimpaired vagal function has been reportedin oesophagitis, we have studied 25unselected patients with IBS (with andwithout oesophagitis) to determine whethervagal dysfunction could explain the fre-

quent occurrence of reflux in this disorder.Rising of intra-abdominal pressure norm-

ally produces an increase in lower oeso-

phageal sphincter pressure through a

vagally mediated mechanism. This response

was subnormal in 13 patients. Efferentvagal function, assessed by the ratio of peakacid output after insulin induced hypo-glycaemia to maximal acid output afterpentagastrin, was subnormal in seven of 23patients. Cardiac vagal function, assessedhy heart rate variability with deep respira-

tion, was abnormal in six of 23 patients. InIBS patients with oesophagitis abnormalityof these tests of vagal function occurredwith a similar frequency to that previouslydocumented in oesophagitis unassociatedwith IBS but abnormal vagal function alsooccurred in patients free from oesophagealdisease. Impaired vagal function is commonin IBS. This explains the frequent occur-

rence of reflux in this disorder and may alsobe important in the pathogenesis of IBSitself.

Increased segmental activity and intra-luminal pressures in the sigmoid colon ofpatients with the irritable bowel syndrome: a

new manometric technique

J ROGERS, J J MISIEWICZ, AND M M HENRY

(Department of Gastroenterology and

Nutrition, Central Middlesex Hospital,London) The pathophysiology of the irrit-able bowel syndrome (IBS) is controversial.We studied seven IBS patients, mean age 51years (31-75) and seven normal controls(NC) 37 years (22-55) with a new techniqueof intraluminal pressure recording in theunprepared true sigmoid colon, using a 4-lumen water perfused tube system placed byflexible sigmoidoscopy at 50, 40, 30, and 15cm from the anus. After a 30 minute basalperiod a 1000 Kcal mixed meal was givenand measurements continued for a further100 minutes. The intubated colon was

treated as a study segment (sum of all fourchannels) for analysis, using Activity Index[AI=area under curve calculated elec-tronically in mm Hg/min] in 10 minuteepochs. Additionally mean amplitudes(MA) per channel were analysed in 10minute epochs. Activity index was signific-antly (0)-001<p<0-05) higher in IBS thanNC in the whole postprandial period; meanAI±SD in the third postprandial epoch was

1425±335 mm Hg/min, IBS v 469±355,NC. Basal Al was similar in IBS and NC.Mean amplitudes in the sigmoid colon were

significantly (0-001<p<0-02) higher post-prandially in IBS than in NC; MA (range) in

the third postprandial epoch was 100 (40)-284) mm Hg, IBS vl 33 (15-257), NC.This study differs from other work in thatpressures were measured in the unpreparedsigmoid colon. Under these physiologicalconditions higher intraluminal pressures

than previously described in normals, andsignificantly higher postprandial colonicactivity in IBS were recorded.

Screening for colorectal cancer - an analysisof 132 tumours

G PYE, K C BAtLILANTYNE, AND J D IHARDCASTLL

(Department of Surgery, UniversityHospital, Nottingham) One hundred andthirty two patients have been diagnosed as

having colorectal cancer within the test andcontrol groups of a faecal occult bloodscreening study for colorectal cancer. In thetest group there were 42 screen detectedtumours (SDT) and 32 tumours in thosewho had refused screening. Ten tumourspresented symptomatically in patients whohad had a negative test. In the control group48 tumours have been diagnosed.Twenty four of 42 (57%) of SDT were

Dukes' stage A compared with 1 1/90 (l12% )of non-screen detected tumours (NSDT),(Xy-=30; p<0()-0l). One of 42 (2-3%) ofSDT were associated with hepatic meta-stases compared with 23/90 (26%) of NSDT(X-= 10; p<0-01). Nine of 42 (21%) of SDTwere pedunculated polyps treated just bycolonoscopic polypectomy compared with2/90 (2,2°%) of NSDT (x2=14; p<0-001).The mean size of SDT was 31 mm comparedwith 50 mm for NSDT (t=4 2; p<0(01).Eleven of 42 (26%/) of SDT were histologic-ally well differentiated compared with 7/69(10%) of NSDT (x2=4-9; p<0-05). Three of42 (7%) of SDT were fixed compared with22/82 (27%) of NSDT (x2=67; p<0.0 1).

Although overall benefit from populationscreening must be judged by longterm mor-

tality, those patients with screen detectedtumours have better prognostic factors.

I IVER POSTERS

Massive haemorrhagic necrosis of the liverafter liver transplantation

S G HUBSCHER, D ADAMS, E ELIAS, J

NEUBERGER, P MCMASTER, AND J BUCKI-LS(Department of Pathology, The MedicalSchool, University of Birmingham,

A1344


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroetnterology

Birmingham) Six of the first 92 patientstransplanted in Birmingham developed asyndrome of fulminant liver failure withcharacteristic clinical and pathologicalfeatures which are presented here. Thetypical clinical profile was of an initiallyuneventful postoperative period followedby a sudden, rapid deterioration in graftfunction resulting in graft failure. All sixpatients died. The characteristic patho-logical changes were those of massivehaemorrhage and hepatocyte necrosis.Initially these changes were confined tocentrilobular areas hut later spread toinvolve entire lobules throughout the liver.Inflammatory changes within necrotic areaswere usually mild. The mode of presenta-tion in our six patients resembles otherpreviously reported cases of fulminant liverfailure after liver transplantation. Themechanisms underlying this process arepoorly understood and a number of aetio-logical factors including ischaemia, infec-tion and rejection have been proposed.None of our cases had any clinical orpathological features to suggest ischaemiaor infection and the changes seen histologic-ally did not resemble other well recognisedpatterns of rejection. A similar picture ofmassive haemorrhagic necrosis has recentlybeen described, however, in an animalmode of hyperacute (humoral-mediated)rejection. The possibility that massivehaemorrhagic necrosis after liver transplan-tation represents a form of acceleratedhumoral rejection is therefore suggested.

Circulating osteocalcin as an index of boneturnover in primary biliary cirrhosis (PBC)pre- and post-liver transplantation

R G P WArsoN, D COUILiON, J A KANIS, P

MCMASTER, AND F, EI IAS (Liver Uniit, QueenElizabeth Hospital, Birmingham, RoyalHallamshire Hospital, Sheffield) Osteo-porosis is a major complication of chroniccholestatic liver disease. Aetiology is notestablished and attempts at treatment havebeen disappointing. Primary biliary cirrho-sis patients have reduced bone turnover(Hepatology 1987; 7: 137-42). In a previousreport, this was associated with reducedcirculating concentrations of osteocalcin(Ann Intern Med 1985; 103: 855-60)) a non-collagen bone protein produced principallyby osteoblasts. In this study osteocalcin wasmeasured by radioimmunoassay in 32women with PBC (ages 31-62 years, mean48) and 52 controls (women aged 35-62years, mean 48). Osteocalcin levels were

lower in PBC's (mean 8±SD 5-6 ng/ml)compared with controls (16-9±5-9)(p<0-00l). In PBC's there waIs nO corre-lation with serum bilirubin, length of historyor history of bone fractures. Levels threemonths to 41/2 years after liver transplanta-tion in 10 patients (including cight in whompre-operative values were available) wereincreased (454+ 19 2) compared with pre-transplant PBC's (p<0(001) and controls(p<() 1(().We have confirmed that circulating

osteocalcin concentrations are significantlydecreased in PBC. After liver transplanta-tion levels are supranormal, suggesting thatbone turnover is increased. This observa-tion lends support to the hypothesis thaithepatic osteodystrophy maiy he arrested orreversed by liver transplantation.

Liver transplantation for fulminant hepaticfailure (FHF)

C VICKERS, J NEUBLRGER, J BUCKELS, PMCMASTER, AN[) E ELIAS (Liver Un1it,Queen Elizabeth Hospital, Edgbaston,Birmingham) The role of transplantation inFHF is uncertain. The transplant unit hasvery little time to assess patient suitabilityfor the procedure and only a short timeperiod exists during which irreversibleneurological complications may supervene.Thus, there is a very short interval to obtaina donor liver. Thirty three consecutivcpatients (13 M), median age 28 (17-64)developed grade III/IV hepatic comabecause of paracetamol overdose (17), viralhepatitis (nine) and other causes (seven).Urgent transplantation was considered in 12(36°%) patents (viral five, other seven).Progressive renal impairment was presentin nine. Our criteria for liver repiacementwas age <55 years, grade 4 hepatic coma fora minimum of 24 hours and deterioratingcoagulation. Cases of deliberate paraceta-mol overdose were excluded. Continuousmedical reassessment was made up to thetime a donor liver became available, with amedian interval of 40 hours (range 1i)-54).Of seven patients transplanted five (71 (0O)

survived. All of these patients are in goodhealth with median follow up eight months(6-20). Three patients died awaiting aIdonor liver and two recovered. Of 21 (64% )patients not considered for transplantation1() (47%) survived, including 54") of para-cetamol overdoses.The results indicate that hepatic replace-

ment may improve survival in selected casesof FHF.

Sulphation and desulphation in normalhuman liver and primary biliary cirrhosis

S IQBAI., C vKVIRS, ANI) F. IllAS (Liver Unlil,Qulee(n Elizahbethl Hospital, Edlgbaiston,Birminghalim) We haive examined the hypo-thesis that cholestasis in PBC is caused byimpairment of' sulphaite ion trainsf'er.Sulphation of bile sailts is importaint tofacilitaite their urinary excretion whereassulphation of sex steroids is important increating inactive hormone. The activities oflithocholic acid (LC) and estronc (El)sulphotransferase and sulphaitaisc innormal livcr (n=7) and PBC (n= 13) wercexamincd. Hepatocytc cytosol wa1s analysedfor sulphotransferasc activity and homog-cnate, prepatred from the same liver, forsulphalt.!sc.

Sulphotransf'erase was significaintly lowcrin PBC compared with controls for both LC(2 63±2()7 v 6 50±3 65, p<0(005) and El(0(33±0+29 v 1-23±0+58, p<0.0005) (nmolproduct formed/mg protein/h, mcan ± SD).El-sulphatasc activity was not signiicantlydifferent from normal liver) 618±+2.65 v'4.37± 184). LC-sulphatase was not detect-able in the one normal and one PBC liverstudied.The results indicate significant deprcssion

of LC and EI sulphotransferase activity inPBC liver. This could causc an accumulationof the hepatotoxic sccondary bilc acid,lithocholic acid, within PBC livcr andimpair conversion of free active cholestaiticoestrogens to their inactive state.

Natural history of primary sclerosingcholangitis (PSC) HLA antigens as predict-ors of prognosis

M L WIl.KINSON, P T I)ONAI DSON, B C PORT-

MANN, J KARANI, ANI) ROGER WIllIAMS(Gastroenterology Unit, Guy's Campus,UMDS oj Guy's and St Thomas' Hosp andLiver Unit, King's College Hosp, London)Survival and natural history of PSC arepoorly understood and prognostic variableshave not been established. We have pre-viously confirmed a positive association ofHLA B8 and DR3 and a negative aissocia-tion of B 12 with PSC and suggested anegative association of DR4. To exatminenatural history, prognostic variables alndtheir relation to HLA antigens the HLAprofiles of 54 N. European PSC patientsundergoing prolonged follow up study werecorrelated with clinical, radiological andlaboratory profiles and with prognosis.Investigations included yearly liver biopsyand ERCP. Poor prognosis (death or liver

A1345


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Societv of Gastroeniterology'transpl.antation) was analysed hy life tables-and correlated with other variables bymultiviariate analysis. Median survival wasovcr 8'/2 years. Poor prognosis was notpredictcd by age, sex, CUC, autoanti-bodies, piecemeal necrosis, prothrombintime, serum albumin, alkaline phosphataseor immunoglobulin (Ig) concentrations.111 A B8 and DR3 were, however, associ-ated with poor prognosis and there wassignificiant linkage disequilibrium betweenthcm compared with 63 laboratory controls.IgM rise (19/45) was associated with HLAA I B8 and/or DR3. Poor prognosis was alsopredicted by serum bilirubin, yGT aind AsT,by histological evidence of cholestasis orcirrhosis and by ERC, pairticularly extra-hepatic appearances.

No specific association between primarybiliary cirrhosis and bacteriuria?

* FIOREANI, 11 MITCHISON, R FREEMAN,M BASSIEND)INE, ANI) 0 JAMI.S (Departments ojMedicinle antd Microbiology, University of'Newcastle Upon Tvnze) This study aimed toconfirm the reported association betweenbactcriuria and PBC, and to examinewhether this susceptibility is related to liverdisease or the 'dry gland' syndrome.

(1) Screening study: we examined oversix months 403 midstream urine (MSU)samples from 160 consecutive unselectedPBC patients (147 F, 13 M), 140 chronicliver disease (CLD) paitients (65 F, 75 M),23 primalry Sjogrens syndrome (SS) patients(,all women). Significant bacteriuria (> 105organisms/ml) was found in 11-2%0 PBC,12l1% CLD and 43/. SS patients (NS).The prevalence of bacteriuria was higher inwomen (12.2%Y PBC and 184% CLD) thanmen (0% PBC and 6-6'Yo CLD). Bacteriuriawas relalted to age (p<()-05 in PBC) and wascommoner in cirrhotics (PBC Stage 1 5- 100;Stage IV 18-1%. CLD; precirrhotic 45%0cirrhotic 14-6%0). E Coli wais the common-est isolate (74%/ PBC, 63% CLD). (2)Prospective study: urine specimens weretested monthly (at one, three, and fivemonths by MSU; two, four, and six monthsby Uricult Dip Slide) from 29 PBC patients- eight 'early', 13 'intermediate', eight 'late'- six CLD and eight SS (all subjectswomen). The cumulative ratte of bacteriuriaof six months was 25' in early, 31%/ inintcrmediate, 50%) in late PBC v 33%/ inCl D, 62-5% in SS (NS). TIhe prevalence ofbacteriuria was higher by dip slide thanMSU (positive dip slides checked by MSUshowed 5/16 ((31 /0) lalse positives) andaccounted for the appatrent higher preval-

ence of bacteriuria in SS. Persistentbacteriuria was found in seven PBC, withthe same organism in 4/7. No CLD patientshowed persistent bacteriuria with the sameorganism.

This study does not support a uniqueassociation between bacteriuria and PBC.SjS6grens syndrome does not appear to be animportant factor for increased susceptibilityto UTI. In both PBC and CLD bacteriuriai ismore common in elderly women withcirrhosis.

Epidemiology of antimitochondrial antibodyseropositivity and primary biliary cirrhosisin the west of Scotland

B M GOUi)IE, G MACFARItANE, P BOYILI., A D

BURI, C R GILItiS, R N M MA(SWi-FN, ANI) G

WAl KINSON (Departments olj Medicine ntidPathology, Westerni Insfirtnary, tanid CancerSurveillance Uniit, Ruchill Hospital,Glasgow, Scotland) During the period1965-1980 715 patients in the West ofScotland (a region with a population ofthree million served by a single immuno-pathology laboratory) were found to beseropositive for antimitochondrial antibody(AMA) using an indirect immunofluoresc-ence technique. Case records have beenreviewed in 647 (90%) and liver histologyexamined in 373 (52%) of these cases. Thenumber of patients satisfying predeterm-ined histological criteria for the diagnosis ofprimary biliary cirrhosis (PBC) was estab-lished for the group in whom liver biopsieswere available and the results extrapolatedto provide an estimate of the frequency ofPBC in the total AMA-positive population.Between 1971-1980 when there was highascertainment of AMA-positive individualsthe incidence of AMA seropositivity was 19per million per year. The estimated incid-ence of PBC was 11-4-15-2 per million peryear and estimated prevalence 69-5-92-7per million. There was sub regional varia-tion in the incidence of AMA seropositivitywith a predominance of cases in urban/industrial areas. The possibility of spacctime clustering was investigated using amodification of Knox's test. No case cluster-ing was shown, the observed number ofpairs close in time and space showing goodagreement with the expected numberassuming no inter.action.

Defective measles virus (MV) antigenexpression causes persistent MV infection inautoimmune chronic active hepatitis(AICAH)

D A F R013[R'i'SON, S 1i ZiIANG, (GUY, ANI)

RAI P1H WRi(iG (I)epartments o0f Medicineanid Microbiology, Southampton GenieralHospital, Solthiamiptoni) Autoimmunechronic active hepatitis (AICAH) is ofunknown cause. High antimeaisles antibodytitres in AICAH suggest a possible role forthis virus, which is capable of persistcletinfection, in incomplete form, in the centralnervous system in subacute sclerosiiigpanencephalitis, despite a very high anti-body response to some mealsles virus (MV)antigens. To investigate the possibility thatpersistent incomplete measles virus infec-tion occurs in AICAH we have measuredcomplement fixing antibodies ((TF)together with Western blot analysis ot theantibody response to meaisles in 18 AICA ItIpaitients (aiuto-antibody positive. bionpsproven), 19 controls with liver disease and16 other disease controls. Very high anti-meiasles antibodies were shown in AICAH(geometric mean CFT 1/512 control 1/32)but these were confined to the nucleocap-sid, cellulair antigens, and polymerase withabsent responses to the glycoproteill.matrix and envelope. These findinigs.together with the demonstration ofsequences from measles virus nucleocapsidgenome in the lymphocytes of 12 of thesepatients suggests that persistent MV inIfcc-tion may play a role in the pathogenesis otAICAH. The correlation between MVgenes encoding viral proteins, ind inti-bodies to the proteins will be described.

Combination chemotherapy in hydatiddisease

D IAYILOR, ) 1. MORRIS, ANI) S RIC'IIARI)S(Departmetnt of Surgers', Universits'Hospital, Nottinlgtiham) Chemotherapy ofEchinoccus remains a diflicult area. Wehave previously reported encouraging invitro ainial and clinicil results withalbendazole and more recently haive shownpraziquaintel to be a very activc scolicidalagent ini vitro. As these two compounds arecompletely different (bcnzimidazole/isoquinoline). their ultrastructural eftectsappear difterent and they do not share thesame toxicity problems, we felt that com-bination chemotherapy should be con-sidered. Combinnation chemotherapy hasnot previously been studied in hydatiddisease. Live ovine protoscoleces (fgrtatinulosis') were maintained in in vitroculture and albendazole (A) 5t)0 ug/l anid/orpraziquantel (PZ) 5), 2)0, l) fIg/l wiasadded with suitable untreated controls.

A1346


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Viability was measured by microscopy/eosin exclusion at three to four day inter-vals. Whilst both agents were active alone,combinations reliably achieved betterrcsults than either drug alone (A v A+PZ50, p<0)-02; PZ 50 v, PZ 50+A 500,p<0).05). Subtherapeutic concentrations ofpraziquantel when given in combinationhad a significant additional effect on viability(A v A+PZ 10, p<0(05). Combinationchemotherapy may allow significantimprovements in efficacy and the avoidanceof dose dependant toxicity problems.

Liver volume, blood flow and perfusion fallin aging man - an explanation for altereddrug metabolism in the elderly

11 WYNNi., 1. 1i COPE, E MUIC'H, M D RAWLINS,K W WOOD)FiOUSE ANI) 0 F W JAMES (Utniversityof Newcastle Upon Tynle) It has hithertobeen assumed that the decreased elim-ination of oxidised drugs in old age in manwas because of agc related decreased specificactivity of human microsomal mono-xygen1asessimilar to that found in rodents;but recent studies have shown no decline inthe spccific activities of several drug meta-bolising enzymes with age in primates orman. We hypothesised that reducedelimination in man might be because ofdecrcased liver sizc, liver blood flow, andliver perfusion with advanced age. Weexamined liver volume by parallel longi-tudinal beta-scan ultrasounds, using themethod of Carr; we calculated liver bloodflow from indocyanine Green clearanceafter iv injection of 0(5 mg/kg. Sixty sixhealthy volunteers (34 women) age 24-91years (20 age <40, 23 age 40-65, 23 age65+) were recruited from hospital staff andlocal social clubs. None had hepatic, renal,respiratory or cardiac disease assessed byclinical history and cxamination. All hadnormal blood count, liver blood tests, nonetook medications. Correlations betweenliver volume, blood flow and perfusion withage were assessed by Kendall's rankcorrelation.

Liver volume fell from 21-4±08- ml/kgbody wt age <40 to 19 3+0-7 age 4(1-65 and155+t)-4 age 65+ (p<0-001). Estimatedvolume was 1470 ml age 24, 945 ml age 91(linear regression). Liver blood flow fellfrom 17 6 ml/min/kg body wt age <40 to15 4 age 40)-64, 11-4 age 65+ (p<0)001).Estimated flow was 1260 ml/min age 24; 655ml/min age 91 (linear regression). Liverperfusion fell from 0(84 ml/min/ml liver age<40) to 0(81 age 40)-64, 0(74 age 65+.

Estimated perfusion was 0-87 ml/min/mlliver age 24, 0-70 age 91. We conclude thatthe age related decline in liver volume,blood flow and perfusion may be the majorcomponent in the liver leading to the faLll inclearance of many drugs in aged man.

IgG3 in PBC: raised serum concentrationsand increased synthesis by peripheral bloodlymphocytes

P BIRD, A FLOREANI, 11 MIICHISON, 0 JAMES,AND M BASSENDINE (Departments ofMedicine and Immuniology, University ofjNewcastle Upon Tyne) Primary biliary cirr-hosis (PBC) is an autoimmune disordercharacterised by positive antimitochondrialantibodies (AMA) and other immuno-logical abnormalities - noteably raisedserum 1gM. We have previously confirmedthat the AMA is predominantly of immuno-globulin subclass IgG3. We have nowexamined the serum concentrations of IgG3in 104 PBC patients (diagnosed on conven-tional biochemical histological and clinicalgrounds), from two separate areas (70English and 34 Italian patients- histologicalStage I, n=27; II, n=14; III, n=28; IV,n=35). A marked rise in serum IgG3 wasobserved in both English (mean 615±63'%serum standard) and Italian (578±96)patients versus age and sex matched controls(mean 70±22) (p<0-001)). Raised IgG3was similar in patients at all histologicalstages. There was, however, a significantcorrelation between serum IgG3 level andAMA titre (p<0(01).To examine the basis of this raised serum

IgG3 (increased synthesis or altered clear-ance) peripheral blood lymphocytes from19 PBC patients and 18 age and sex matchedcontrols were cultured in vitro and theimmunoglobulin synthesised in the super-natant was measured by capture ELISAspecific for total IgG, IgG3 and IgG3 Kappaand Lambda chains. Primary biliary cirr-hosis lymphocytes in culture spontaneouslysynthesised a higher % IgG3/total IgG thancontrols (p<0(001). This correlated withtheir serum IgG3 (r=0-69, p<0-001) andindicated that there is increased synthesis ofthis isotype by lymphocytes in PBC. Noevidence for clonally restricted synthesis oflgG3 by PBC lymphocytes was found onanalysis of the IgG3 Kappa/Lambda ratio.Some viruses preferentially elicit IgG3 anti-viral responses; conceivably identificationof factors which trigger this increased syn-thesis of IgG3 in PBC may provide clues tothe underlying pathogenesis of the disease.

Haem arginate therapy for acute hepaticporphyria

K E L MCCOiL, M R MOORE, A III.RRI(K, M .1

BRODIE, AND A GOLD)BEIRG (UniversityDepartmenit of' Medicine, WesterniItnfirmary, Glasgowu) Thc acute hepattic por-phyrias are causcd by hereditary deficicn-cies of individual enzymes in the pathway ofhaem biosynthesis. Consequently, there isincreased activity of the initial rate control-ling enzyme of haem synthesis dclta-aminolaevulinic acid (ALA) synthase, overproduction of the haem precursors formedbefore the enzyme block and impaircdhepatic mono-oxygenase activity rclalted toinadequate hepatic haem synthesis for cyto-chrome P450 formation. We have studiedthe effect of a new haem preparation, haemarginate (3 mg/kg/24 h iv for three days) in20 attacks of acute porphyria in six patients.The mean urinary excretion of the haemprecursor ALA pretreatment was 32(1[tmol/24 h (range 180-602) and this fell towithin the normal range (0-40) in each caseby the third day of theratpy (p<0(001).Leucocyte ALA synthase activity was 120(1nmol ALA/g protein/h (range 530-1800)before therapy and fell to 489 (range 215-720) 24 h after the last infusion. Antipyrinet'2 (an index of mono-oxygenase activity)was 18-7 h (range 7-2-41) over the 48 himmediately before starting therapy and fellto 8-7 h (range 4-12-4) during the first 48 hof treatment (p<0,- 1).These results confirm that haem arginate

effectively corrects the biochemical mani-festations of acute hepatic porphyria and issuperior to previously used therapics. Thedemonstration that impaired hepatic mono-oxygenase activity can be corrected bygiving haem arginate indicates that theexogenous haem is being incorporated intofunctionally active hepatic haeemoproteins.

Nutritional support in fulminant hepaticfailure: the safety of lipid solutions

AlASTAIR FORBES, CLAIRE WICKS, WIll1IAMMARSHiALL, PHILIP JOHINSON, PAUI. FORSEIY,AND ROGER WILLIAMS (King's CollegeHospital and School of Mediciine antdDentistry, Denmark Hill, London) Whenfulminant hepatic failure (FHF) is pro-longed the necessary nutritional support ishampered by concomitant renal failure(limiting possible volumes), by insulinresistance, and by a theoretical needto avoid lipid solutions (risk of toxicityfrom fat accumulation and by thrombo-cytopenia). In the present study the

A1347


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society ofGastroenterology

magnitude of the risk of iv lipid adminis-tration has been assessed by acute lipidchallenge, and by prolonged lipid infusionas part of a parenteral feeding regime. Ninepatients with FHF complicated by renalfailure and cerebral oedemra (from para-cetamol intoxication in eight from hepatitisA in one were given 500 ml Intralipid 1(%over four hours. The median basal tri-glyceride level was 0(7 mM (0-4-1-7), thatat the end of infusion 12 0 (7-1-22-5), aindwithin eight hours all had levels -2-0. Therewere no complications and platelet countswere unaffected. Four patients withanthropometric evidence of millnutritionwent on to receive TPN which included 500ml Intria lipid 2("0% daily, for 7, 1(1, 15, and 41days respectively. No complications attri-butable to the fat content were observed,liver function tests continued to improve,aind serial triglyceride levels did not riseabove 1-5 mM. All four pattients made a fullrecovery and it is concluded that lipid solu-tions may safely be incorporated into nutri-tional support for patients with FHF.

Carbohydrate meals increase hepatic venouspressure gradient (HVPG) in cirrhotics whohave bled from oesophageal varices

P A MCCORMACK, M GRA-FFEO, R D)ICK, N

MORREAILE, I) WAGSIAFF, A MADDiN, N

MCINIYRI., AND A K BURROUGHS (Royal FreeHospital School of Medicine, London)Protein but not carbohydrate or fat feedingcauscs a marked rise in hepatic blood flow innormal subjects and protein also signitic-antly increases (HVPG) in cirrhotics. Weexamined the effects of a protein free meal(19 g carbohydrate, 5 g fat and 120 Kcal per100 ml) on HVPG in 16 cirrhotics (15 men:median age 49 range 20)-56: alcoholic n= 10O,cryptogenic n=5, aiutoimmune n=i) whohad bled from oesophageal varices. Eightreceived a 600 Kcal (500 ml) and eight a30)0 Kcal (250) ml) meal. Pressures weremeasured before and at intervals of tenminutes for one hour after the meal.Hepatic venous pressure gradient increasedin all patients after 600 Kcal rising 29%o from15-6±SD 5-1 mm Hg to 20-2±SD 6-2 mmHg at 30 minutes (p<0.)l) and by 22%to 19 1+SD 6-0 mm Hg at 60 minutes(p<0.0l). After 300 Kcal HVPG increasedin 7/8 patients, mean HVPG increased from185±SD 7-7 mm Hg by 15% to 21-2±7-1mm Hg at 30) min (NS) and by 6% to19-6+8-2 mm Hg at 60 minutes (NS). Theseresults suggest that despite the flow resultsin normal subjects non-protein mealssignificantly raise portal venous pressure in

cirrhotics with varices and that there is adose effect which may be of importance inthe management of patients following acutevariceal haemorrhage and for the longtermadministration of portal hypotensive drugs.

Prospective randomised trial of endoscopicsclerotherapy (ES) and transhepatic emboli-sation (TE) in patients with severe cirrhosisand acute variceal bleeding. Preliminaryresults

P BONNIERE, J F C(LOMBELi, A CORTO-Il, M

IIIFNRY-AMAR, C 1. IiFRMINE, AND J C PARIS

(Hepato-Gastroenterology and RadiologyDepartments, University of Lille II, Lilleancd Biostatistics Department, Villejuif,France) Seventy five Child's C bleedingcirrhotic patients have been randomised toES (n =25), TE (n= 25) or to a control group(C) (n =25). Initial patient's characteristicsdid not differ within the three groups especi-ally for clinical status and bleeding.The mean time between admission and

treatment was 8-3 h in ES patients, 8-2 h inTE patients, and 6-9 in C patients (NS).(1) A bleeding control was achieved in 25/25(1()()%) ES patients, 18/25 (750%) TEpatients, and 19/25 (76%o) C patients(p=0.02). (2) The mean number of bloodunits required two days after randomisationwas 0(76 in ES patients, 18 in TE patients,and 2-4 in C patients (p=0-02). (3) An earlyrebleeding during hospital stay occured in3/24 (12%) ES patients, 1/16 (6%) TEpatients, and 5/17 (29(%) C patients (NS).(4) Eight patients died (32%) in ES group,11 patients (44% ) in TE group, and 11patients (44%) in C group during the sameperiod (NS).We conclude that (1) ES was better than

TE and usual medical treatments for stop-ping active variceal bleeding; moreover, itsignificantly reduced blood requirements.(2) Although lower in ES, early death ratewas not significantly improved by ES or TEas compared with control group.

Is bleeding time a useful test in cirrhoticpatients?

D SPRENGERS, P (RECIIl, P A MCCORMICK, N

MCINTYRE, AND A K BURROUGHS (Royal FreeHospital School of Medicine, London)Bleeding time (BT), a good test for primaryhaemostasis, is not routinely done in cirr-hotics. Bleeding time and platelet count(PC) correlate directly below lOOx 109/l inhypersplenic thrombocytopenia. Increasedbleeding risk is estimated with prothrombin

time (PT) , 17s and/or PC <80x 109/I, butno correlation with BT has been made. In100 cirrhotics PC, PT, partial prothrombintime (PTT), and BT in minutes (simplate 11General Diagnostics) were studied. Fortytwo had a BT ¢ 10. Bleeding time correlatedwith PC, above and below IOOx 10)/I(Spearman's p<0-(001). Of 63 cirrhoticswithPC 80x10)9/1, 19(30%/0)hadaBT>l0Iand 10 (16%) had also a PT <17s. Of 37with a PC <8)x 109/1, 17 (380/.) had anormal BT and nine (14%) had also a PT¢ 17s. Forty five cirrhotics had a PT ¢ 17s ofwhom 22 (53%) had a normal BT; 55 hadPT <17s of whom 22 (4(0%) had a BT >10.Cirrhotics with BT _t1), had worse liverfunction, without correlation between BTIand PT, PTT or PC, suggesting that factorsother than thrombocytopenia prolong BT.This is the first report of a statistical correla-tion between PC and BT in cirrhosis. A PC>80X 109/1 and/or a PT < 17s, however, donot predict a normal BT and vice versa. Asplasma fraction iproducts do not correctabnormal BT and platelets are not usuallygiven for PC ¢2(80x 109/l, an unrecognisedabnormal BT may remain uncorrected.

Metallothionein and copper in liver diseasewith copper retention: a histopathologicalstudy

M E, ELMES, N J MAHY, J P C(LARKSON, AND)

B JASANI (Department oJ Pathology,University of Wales College of Medicine,Heath Park, Cardiff) Histochemical studiesof copper and copper associated protein(CAP) correlate poorly with hepatic coppercontent in Wilson's disease (WD). Byexamining immunoreactive metall-lothionein (IMT) immunocytochemicallyand Cull and CAP using rubeanic acid andorcein respectively in 120 diagnostic liverbiopsies we attempt to explain this discrep-ancy and investigate the diagnostic value ofMT immunocytochemistry.

In normal liver (n=25) Cull and CAPwere absent and cytoplasmic IMT waspresent in centrilobular hepatocytes. In pri-mary biliary cirrhosis (n= 14) Cull and CAPwere often found in paraseptal hepatocytesand IMT had a strong overall distribution.Degenerate hepatocytes in piecemnealnecrosis were intensely IMT positive. Inchronic active hepatitis (n=25) with piece-meal necrosis IMT positive cells wereprominent against pale IMT staining ofbackground hepatocytes in contrast withWD (n=5) where all hepatocytes werestrongly IMT positive. In sclerosingcholangitis (n=4) and nonspecific cirrhosis

A 1 348


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


(n=47) with cholestasis CulI and CAP werefound but IMT had no consistent pattern.We suggest that metallothionein bound

copper (Cul) which is not demonstratedhistochemically is detected by MT immuno-cytochemistry. This explains the discrep-ancy described and will be helpful inseparating piecemeal necrosis caused byWD from other causes.

Mitochondrial aspartate aminotransferase(mAST) as a marker for alcohol misuse inpatients with liver disease and controls

C G ALVEYN, J BERESFORD, F G BULI, AND

RAILPH WRIGHT (University of Southampton,Southampton) Eighty three patients (51men and 32 women) were studied, 22 withalcoholic liver disease, 20 alcoholics withnormal liver function tests, 26 with non-alcoholic liver disease, 15 epileptics onanticonvulsants; and 64 normal volunteers(32 M and 32 F) before and after ethanolchallenge (1.1 g/k daily for one week).Alcohol consumption was assessed byquestionnaire. Blood samples were takenfor routine liver function tests (LFT's),mean red cell volume (MCV), glutamyltranspeptidase (yGT), and mAST by themethod of Rej. the ratio of mAST to totalAST (tAST) was calculated.

Significant differences were observed inthe mAST/tAST ratio, the highest levels inalcoholics currently drinking heavily(median mAST/tAST ratio 21.6%). Thosewho were drinking moderately (<80 g/day)or had stopped drinking were significantlylower (14.6% and 11 8% respectively). Theratio was significantly lower (median9.45%) in patients with non-alcoholic liverdisease and significantly elevated in epilep-tic patients on anticonvulsant treatment(median 15.2%). In normal subjects, therewas no difference in the mAST/tAST ratioafter one week's alcohol challenge.When compared with the yGT and MCV

the mAST/tAST ratio was a better index ofcurrent heavy alcohol consumption thaneither of these standard parameters.

Glutathione S-transferase levels in auto-immune chronic active hepatitis

P C HAYES, A J HUSSEY, J KEATING, I A D

BOUCIJIER, R WILLIAMS, G J BECKETT, AND J D

LIAYES (University Departments of Medicineand Clinical Chemistry, Royal Infirmary,Edinburgh and Liver Unit, King's CollegeHospital, Denmark Hill, London) Serumglutathione S-transferase (GST), a sensitive

marker of hepatocellular damage, wasmeasured by radioimmunoassay in 26serum samples from 22 patients withhistologically proven, autoimmune chronicactive hepatitis (CAH) and compared withaspartate transaminase (AST) levels atvarious stages of the disease. All but onepatient were on steroid therapy and 13 werealso receiving azathioprine. In 17 of the 26paired results, GST was raised comparedwith only six AST levels. In 11 samples,an abnormal GST was associated with anormal AST concentration. All AST levelsin patients with GST levels below 8 ng/l (twotimes upper limit of normal) were normaland only once when this threshold value ofGST was exceeded was there a correlationbetween GST and AST. Patients with histo-logically active CAH demonstratedabnormal levels of both enzymes but manypatients with chronic persistent hepatitisand little evidence of activity still hadabnormal GST levels.These data show that ongoing hepato-

cellular damage is occurring in patientstraditionally considered in biochemicalremission and may explain why manypatients with treated CAH progress to cirr-hosis despite treatment.

Effect of high and low protein meals onportal haemodynamics in patients with cirr-hosis and portal hypertension

N H AFDHAL, P A MCCORMACK, M KEOGHAN,AND J F HEGARTY (Gastroenterology & LiverUnit, St Vincent's Hospital & UniversityCollege, Dublin) Although splanchnicblood flow increases after a meal in normalsubjects, the postprandial increases inhepatic blood flow and its relationship tochanges in portal pressure in patients withcirrhosis and portal hypertension have notbeen extensively studied. In this presentstudy the portal pressure (PSG; wedgedhepatic venous pressure minus free hepaticvenous pressure) and estimated hepaticblood flow (EHBF), calculated using Indo-cyanine green clearance, was determined innine patients with cirrhosis and oesophagealvarices following a 20 g and 80 g isocaloric,isovolumetric protein meal.

After the 20 g protein meal PSGincreased by 22% from 11-6±SD 6-3 mmHg to 14 2±6-6 mm Hg (Student's t,p<O-01). After the 80 g meal PSG increasedby 33% from 14 1±SD 3-9 mm Hgto 18 7±SD 3-8 mm Hg (p<0001). The risein PSG after the 80 g meal (47±+SD 1-7 mmHg) was significantly greater than thatoccurring after the 20 g meal (2.5+±22 mm

Hg; p<0'05). Maximal increases in PSGoccurred at a median time of 20 minutesafter the meal and were sustained for aperiod of observation of up to one hour.The EHBF after the 80 g protein mealincreased by 56% from a fasting level of1399±SD 207 to 2474±SD 215 ml minute.Increases in EHBF occurred within 20minutes of the meal and coincided with theincreases in PSG.These studies indicate that in patients

with portal hypertension protein mealscause significant changes in portal haemo-dynamics that are dose dependent andrelated to increases in hepatic blood flow.

Is the reduced vascular tone in cirrhosiscaused by 'desensitisation' to sympatheticnervous activity?

A J MACGILCHRIST, D J SUMNER, J L REID, AND

r J THOMSON (University Departmentof Materia Medica, Stobhill Hospital,Glasgow) Severe cirrhosis is associated withreduced peripheral vascular tone despitesympathetic overactivity. We have investi-gated whether sympathetic desensitisationmay explain this anomaly. We measuredblood pressure (BP) responses to steadystate intravenous infusions of the sympath-etic agonist noradrenaline (NA) and thenon-sympathetic pressor agent angiotensinII (All) in 10 severe cirrhotics (Pugh'sGrade B or C, group 1), 10 milk cirrhotics(Pugh's A, group 2) and 20 healthy controls(group 3), matched for age and alcoholintake. Before the study, the severecirrhotics had raised plasma NA andlowered BP compared with controls. Thedose of NA or All which raised BP by20 mmHg (PD2,,) was calculated fromindividual dose response curves. Figures aremean±SD, comparison by one-wayANOVA after log transformation.

Group Group2 (roup3 v 3PD210NA(4g/kg/min) 0)13+1)±1)0 1)1)7+±01-)2 01X+I±)0 1)p<) 05PD211 AII (ngkg/min) 2)) 6±21 9 6 ±+44 6 1l+ 8 p<0) l)2

Thus there is reduced vascular responsive-ness to both sympathetic and non-sympathetic pressor agents in severecirrhosis. We conclude that sympatheticdesensitisation may indeed occur in thesepatients, but is not the sole explanation forthe loss of vascular tone.

Effect of obstructive and hepatocellularjaundice on triglyceride metabolism

ISABEL-MARTINEZ, R I HALL, G R GILES, ANDM S LOSOWSKY (University Department of

A1349


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Surgery, St Jamnes'Xs University Hospital,Leeds) Jaundiced patienits exhibit at numberof abnormalities of lipid metabolisms.These may have important implicationswhen they require nutritional support.Plasma triglyceride (Tg) turnover was

studied in patients with obstructive jaundice(OJ), hepatocellular jaundice (FICJ) andcontrols with normal liver function. Paitientsreceived an intravenous bolus of 1-2-glycerol after pliasmalt Tg concentration was

confirmed to be in steady statc. The deciaycurve of pliisma radiolabelled triglycerideilctivity was used to obtain the fractionalcatabolic raite (FCR) and total Tg turnover

(V).Pliasma I g turnover was reduced almost

10-fold in patients with HWJ (1-9±0 7mg/kg/h) compaircd with cointrols ( 18()+2-2 mg/kg/h) or paitients with OJ (1933+3-7mg/kg/h). The ratios of plasmaiIag turnover/concentration suggest that patirents with OJhave reduced plasma clearance capatcity butnormal hepatic Tg secretion. Patients withHCJ have reductions in both hepatic Tg

secretion and plasmna clearance.The results indicate that patients with

HCJ and OJ may haive impLired ability toclear artificial fat emulsions from theplasma. Patients with HCJ maty be at riskof developing hepaitic steiatosis duringnutritional supplemenitations due to theirreduced hepatic Tg secretion.

Interstitial hyperthermia - a new therapyfor intrahepatic neoplasms

A C STEGEIR, K MATIIIEWSON, S ( BoWN, AND CG('I ARK (Nationatl Medical Laser Centre,Department o' Siurgery, Univer.sity College,Loncdon, The Ravne lntstitlute, Lonidoni) Alow power laser coupled to a thin flexiblefibre, inserted directly into tissue (inter-stitial therapy) can deliver energy to thecentre of solid organs with great precisionand without damage to overlying tissue.This produces local thermal damage whichis predictable in nature and extent.

In normal rats we established that themaximumil diameter of necrosis producedwith a single interstitial fibre was 18 cm,

with I Watt ot power tor 600s from a ND-YAG laser. (Below I W less necrosis

occurred; above 1.5 W severe charring andvapourisation occurred around the fibretip.) Healing was by regeneration. withminimal scarring within six weeks.

I arger volumes of necrosis were pro-duced in normal canine liver using multipletreatment sites with four fibres activatedconsecutively or simultianeously from the

same laser. Maximum necrosis (7x4 cms)was (achieved with 15 W for 67ff secs downeLch fibre simultaneously. These lesions

healed safely, mainly by regeneration, withlittle systemic upset.

Larger volumes could be treated usingmore fibre insertionsi. Thermally inducednecrosis has a potential for treating livertumours and healing is likely to occur in a

similar way to necrosis in normanl liver. Thispossible new treatment for ultriasonicallydefincd intrahepiatic neopliasms, shouldcarry a lower morbidity than chemo-therapy, partial hepatectomy or liver trans-plantation.

EfTect of SMS 201-995 on the growth anddevelopment of hepatic metastases

I) M NOJI, S A JENKINS, S (iRIMI, J YAIES, D W

[)AY, .I N BAXII.K, AND (i COOKI (UniiversityDepartments of' Surgery, Pathology antdNuclear Medicine, Royval LiverpoolHospital, Liverpool) As reticuloendothelialsystem (RES) inhibitors promote tumourgrowth the present study was carried out toestablish whether an RES stimulant, SMS201-995, had any effect on the growth anddevelopment of liver metastases. Hepaticmctaistascs werc induced in Fisher rats by

intraportal injection of 4x 10( Walker cells.An cxperimental group (n=20) received2 Ftg SMS 201-995 sc bd commencing on

the day of inoculation. Control animals(n=20) werc dosed similarly with isotonicsalinc. All rats were killed three weeks laterand the extent of liver replacement by

tumour assessed. Hepatic and splenic RESactivity was determined by the uptake of"9Technetium sulphur colloid and the effectsof SMS on Walker cell growth measuredboth inl vitro and ini vivo. SMS 201-995significantly inhibited the growth anddevelopment of hepatic metastases (p<)-()5Mann Whitney U Test). Thus in twoanimals there was no tumour and in a

further 12 solitary metastases (1-2 cm

diameter) were present. In four rats treatedwith SMS 25% of the liver was replaced by

tumour. In contrast in control rats thedegrec of liver replacement by tumourvaried between 60 -90/0. SMS stimulatedthe growth of Walker cells ini vitro but hadno eftect ini vivo. RES activity was stimu-lated by SMS, this effect being particularlymarked in the liver (liver: blood "Tc-ScSMS 4-41 +0-45, controls 2-20±0(41p<0t)2, Mann-Whitney). These resultsclearly indicate that SMS 201-995 signific-antly inhibit the growth and development ofexperimentally derived hepatic metastascs,possibly through RES stimulation.

SMAILI INI'ESTINE POSII'ERS

An assay for folate catabolism - implicationsfor the RDA

MG COURINEY, J M MCPARILIN,J M SCOII, ANI)* (G WI IR (Dept of Clinical Medicine, SiJames's Hospital and Trinity College,Dlublimi) The assessment of folate status bylaboratory based criteria is inaccurate,costly and invasive. No measure of folaiteutilisation in vivo or reliable folate balancestudics exist. Using radioisotopes we haveshown thiat cellulair folatte catabolisminvolvcs cleavage at the C9-N 1t) bond withsubsequent urinary excretion or para-aminobenzoylglutamrate (pABGlu) and itsacetylated derivative. We hiave developed amethod to quantitatively estimnate theselatter two compounds, without the use olfaradiolabel in 24 hour urine collections inanimails and man. The assay involves puri-fication by ion exchange chromatographybefore diazotisation and polyacrylamide gelchromatography prior to analysis by highpressure liquid chromatography (HPLIC).Daily urinary excretion (~tg/day) ofpABGlu in healthy women was 28-2+9-1(x±SD, n=58) and in healthy men was64 6±28X4 (n=42). As PABGlu representsone half of the degraded folate molecule,these results indicate a daily folate catabo-lism of 56 sg for these women and 129 ,tgthese men.These excretion derived values approxi-

mate 15% and 30(% of the recommendeddietary allowance (RDA) of folate forwomen and men respectively, suggesting asubstantial overestimation in the RDA.

Incorporating lipid into soup or main coursehas different effects on gastric emptying andnutrient absorption

K M CUNNINGHAM AND N W REA[) (Sub-Depftof Humnan Gastrointestinal Physiology, atidNuitritioni, Uniiversity of Sheffield, Sheffield)Three studies were carried out in each of sixvolunteers to investigate how lipid, given atdifferent stages during the course of a meal,affects gastric emptying and postprandiillblood glucose levels. The meal consisted of300 ml low nutrient soup, followed 20minutes later by a 300 g mashed potato.Sixty grams margarine was incorporatedinto either the soup or the mashed potato.Incorporation of fat into the soup delayedgastric emptying of the mashed potato com-pared with the fat free control meal (ti-=114+11 v 67±7 min; p<0.05) by increasingthe lag phase (31±7 v 2±1 min; p<0-05),

A 1 350)


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


but not affecting the slope (0-9±0.1 v0-8±0Xl"/ /min). This was associated with adelay (1 15+13.8 v 57 5+3-3 min; p<0(05)and a reduction of the glucose peak(75+5±-5 v 9-1±0-5 mmol/l; p<0-05).Incorporating fat into the mashed potatoalso slowed gastric emptying (ti,,= 102±9 v67±7 min; p<0-05), but by reducing theslope of emptying (0(9±0+1 v 0(6±0+1%/min: p<0-05), not by altering the lag phase(4± 1 i 2± 1 min). This was associated with agreater reduction, but no delay in the peakglucose level (6-6+0-4 mmol/1) than withthe high fat soup. Our results indicate thatthe effect of fat on gastric emptying andabsorption of nutrients depends upon atwhat stage of the meal it is eaten.

Evaluation of the nutritional effects andsurgical risks during a two year study aftergastroplasty for morbid obesity

N 1) (ARR, R IIARRISON, R BAUGIHEN. A

1OMKINS, AND C G CI ARK (Department ofSuirgery, University College, Lon7don, TheRaYne Institite, London, and DepartmentofJ Human Nutrition, LSHTM, Lonidoni)Vertical banded gastroplasty (VBG) hasreplaced jejuno ileal bypass in the surgicaltrcatment of morbid obesity. Although thisprocedure carries a low incidence of earlypost-operative complications, longer termfollow up and the impact on nutritionalstatus have not been described.

Forty five patients with morbid obesity[mean body weight 128-3±22-1 kg; meanbody mass index (weight/height) 46-8±6-5kg/m'; mean age 42-2±8-6 years] under-went VBG and were followed up in a specialclinic. The accumulative weight loss was(mean±1 SD) 26-2+9 3; 33X8±13 7 and39-7±16-4 kg at 6, 12, and 24 monthsrcspectively. Body mass index (mean+ ISD) fell to 37-8±5-8; 35-5+6-2 (and 33-1 +5-4 kg/m' at 6, 12, and 24 months post-operatively. Immediate surgical complica-tions were minimal but during follow up,revision of gastroplasty was required in fivepatients because of' severe vomiting. Themajority of patients, however, were aible toeat a normal diet in reduced quantities andexperienced great improvement in thequality of life.The benefits and relative safety of VBG

suggest an important role for this procedurein the management of morbid obesity.

Do we need to regrade postacute gastro-enteritis?

I) A KEILiYx, J A ARMIiSTiEAD). ANI) A WAI KI.R-

SMi rii (Quleen Elizabeth lospf)italChildreni, Lonldoni) Sixty eight ba.bies undernine months were studied in order to evailu-imte current feedting praictices. All babies

were fed glucose electrolytc mixturc(GEM) for 24 hours and theni randomlyassigned to three groups: group I (n=22)were regraded onto a low solute milkformula (ISM) over four days; group 11(n=22) were fed full strength LSM andgroup III (n=24) were rapidly regradedonto a hypoallergenic whey hydrolysateformula. Babics were followed for five dlaysand reviewed at 14 days. Groups were

evenly matched lor age, sex, ethnic origin.nutritional statc and dcgree of dehydration.There was no significant differencc in

vomiting, stool frequency. presence of

reducing substances, or duration of hospitalstay between the groups. [here was poorcompliance duc to refusal of feeds in thebabies fed the whey hydrolysiatc only (group111). Weight was more rapidly regained inbabies in group 11 although weight gain at 14days was comparable in all groups. Twelve(17'%0) babies relalpsed (group 1=4. group11=5 group 111=3). An entcric pathogenwas detected in 9/12 and cow's milk proteinintolerance was diagnosed on jejunal biopsyin the remaining 3/12 (one from eachgroup). In this group of previously healthy.well nourished babies with mild acutegastroenteritis there is no advantage in a

slow regrade to milk feeds or in fecding a

hypoallergenic formula.

Comparative value of tests for smallintestinal bacterial overgrowth (SIBO)

I) IF LOF-F, S A RILEY, ANi) M N MARSII

(Department of Medicine, Hope Hospital(University of Manchester School ofMedicine), Salfrord) Culture of smallintestinal contents is generally regarded as

the 'gold standard' for diagnosis of SIBO.Many less invasive tests are available butlittle is known of their comparative value.We have studied, prospectively, the sensi-tivity and specificity of: ( I ) jejunal aspirateculture (JC); (2) urinary indicans excretion(Ul); (3) glucose hydrogen breath test(GH); and (4) '4C-glycocholate breath test(GC). Thirty four subjects (13 F:21 M)underwent JC, 24 hour Ul excretion andsimultaneous 50) g GH and GC breath test.Symptomatic response to treatment was

assessed following either oral metronida-zole or tetracycline. Sixteen of 34 patientshatd SIBO. Predisposing causes were foundin 13.

( I ) JC true positive 13/16 (senlsitivity81"/<o). 0/18 fllsc positive (specilicity 100().(2) Ul true positive 8/15 (sensitivity 53/0).false positive 3/17 (specificity 82%).(3a) (Afl true positive 13/16 (senisitivity81Io). fallse positive 3/ 18 (speciticity 83' ).(31) GC true positive 10/16 (senisitivity62.5%') fl'tsc positive 3/18 (speciticity 83XYo).(3c) Combined Gll and GC true positive14/16 (sensitivity 87-5%). false positive12/ 18 (speciticity 66.6'6,).

In the diagnosis of SIBO: ( I) false neca-tive JC is not uncommon; (2) (iH hatssimilar seisitivity to 1(; (3) combined Gtiand GC results in minimal incrcase in sensi-tivity with loss of specificityv (4) (GC ind UlIhave unacceptably low sensitivities.

Human serum IgA response to (Giardialamblia

* K J (ioKA, P) 1) K R)OLSiON, V I MA'I'IIAN, ANI)

* J (i FARTH'IING (Deplt (,aZsiroewel(roslo,¢Y,St Bartioloinevs7o.sp0), a11(1 WellchmieResearclh Unit. (CM( llo.spital, Vellore.lnd(la) Specific scruIml IgG and1gMresponses occur in giardilsis. the lItterbeing helplul in diagnosis. Using El-ISA wehave now measured specific IgA rcsponisesin 78 paitients with diarrhoea in the UK andIndia. 39 ol whoni haid giardiasis. (Giardilsiswas excluded by multiple stool cxarmina-tioiis and in Indian patients by duodenalaspirate microscopy. Antigiiardia IgA titreswere higher in control subjects from India(medianit10)0 range 0)-20)0)) thain in thosefrom UK (0,) 1-(1()). IgA titres for Indianand UK patients with ( lambaia were 20)0)(0)-1600) and 50) (0)-3200(). rcspectively.Nine of 25 (36%) ot UK and lour of 14(28% ) of Indian patients with giardilsis haIdraised IgA titres. Eleveni of 39 (28%) ofIndian and UK patients with gi.irdiaiisis hadno detectable anti-Giardlia IgA. therc bcingno difference in the proportion of IgA-negative paitients from cither geographiclocation. Ot the fivc patients on whomconvalescencc serai were Cavailahle two tofour weeks aifter pairaisite eradication. IgAtitres decreased in two, were unchainged intwo and rose in one.Thus spccitic serum IgA responses occur

in giardiasis in approximately one-third ofinfected individuals. High scrum IgA titres(>2(0)0) were never found in either InldiaIn orUK disease controls, suggesting thit thisserological test may hc of value in diagnosiswhen titres are raised, bLut low titres do notexclude infection.

A 1351


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Intensive dietary counselling in adolescentswith cystic fibrosis

J WILLIAMS, D A HANDY, P H WELLER, AND I W

BOOTH (Institute of Child Health, Universityof Birmingham and Birmingham Children'sHospital, Birmingham) Increasing energyintake in patients with CF has been shownto have variable results; failure is frequentlyassociated with poor compliance. Twentyeight adolescents with CF (9-18; mean:13.7) were studied over one year to assessthe effects -of intensive dietary counsel-ling upon nutritional status. They wererandomly allocated to three groups: group(A) controls; group (B) received counsel-ling to improve dietary intake to >130% ofRDI (expected weight for height) usingnormal foods and simple supplements;group (C) received counselling and also anenteric coated pancreatic enzyme supple-ment instead of conventional supplements.The groups were matched for age, sex, andShwachman score.

Surprisingly, the only group whichshowed a significant change in weight whencompared with the other groups was groupB (mean weight SDS change±standarddeviation: Gp (A) -0-143+0 214; Gp (B)+0-037+0-245; Gp (C) -0-131±0-176(p<0-001)). Although there was also anincrease in weight velocity in this groupcompared with groups A and C, this did notreach statistical significance.These data suggest that intensive dietary

counselling may be more important thanenteric coated pancreatic enzyme supple-ments in improving nutritional status inadolescents with CF.

Effect of ritiometan on food pseudo allergy

C ANDRE, F ANDRE, AND L DESCOS

(Laboratoire d'immunopathologieDigestive INSERM Centre HospitalierLyon-Sud, Pierre-Benite, France) Digestiveand extra digestive symptoms caused byfood pseudo allergy are identical to those ofan anaphylactic reaction but are caused bynon-specific mechanisms. Gut inflamma-tion may be an important factor in foodpseudo allergy. Ritiometan (INN) reducedeosinophil chemotactic leucotriene genera-tion and histamine release from mast cells;it reduced PAF synthesis and inhibited C3activation. Its effectiveness was thereforeassessed in seven patients presentingtroubles after the consumption of tyraminerich or histamine releasing foods. In thesepatients prick tests to food were sometimespositive and related to non-specific hista-

mine release but food specific IgE measure-ments were always negative as well as thedetection of jejunal IgE plasma cells.

Gastrointestinal permeability was invest-igated by the ingestion of 65 ml watercontaining 5 g lactulose, as a marker ofabsorption of large molecules and 5 gmannitol as a marker of absorption of smallmolecules. Complete five hour urine collec-tions were made, volumes recorded andaliquot of each preserved for sugar analysisusing gas chromatography.

In 100 healthy subjects mean mannitolexcretion was 14.3%0 (SD+3.3) and meanlactulose excretion was 0(30(% (SD±+020).In the seven food pseudo allergic patientsthe mean mannitol excretion was 16.0(%(SD+5.4) and the mean lactulose excretionwas 1-06% (SD+±055).

Oral ritiometan (150() mg daily for a week)was associated with clinical improvement.This treatment resulted in a significantincrease of mannitol recovery (27-5o;SD+6.6) and in a significant decrease oflactulose recovery (0 62%; SD±+047).We concluded that raised gut perme-

ability to large molecules, probablv relatedto inflammation mediators release, may beinvolved in the pathogenesis of food pseudoallergy. These abnormalities respond to theoral treatment with ritiometan.

Immunofluorescent and genetic studies ofcoeliac relatives of patients with dermatitisherpetiformis (DH): does sub-clinical DHexist?

S O MAHONY AND M J WHELJON (Departmentof Gastroenterology, Regional Hospital,Cork, Ireland) Two previous studies foundno IgA skin deposits in normal relatives ofpatients with dermatitis herpetiformis(DH). We have studied a distinct subgroupof relatives of patients with DH, namelythose with coeliac disease (CD). We havepreviously reported on 119 patients withDH, 13 (1177%) of whom had 21 relativeswith a clinical history suggestive of CD. Wehave established a histological diagnosis inseven of these and have studied skin IgAand HLA typing in this subgroup. All wereon gluten free diets. Two had IgA depositsin the skin and one of these had DH. Sixwere HLA B8 positive, five were HLA Alpositive and all five who had DR studieswere DR3 positive. The six probands all hadIgA deposits in uninvolved skin and five ofthe six were B8 positive, four were Alpositive and of three studied two were DR3positive. The genetic homogenicitv and thepresence of IgA in the skin in some of the

coeliac patients suggests that CD and DHare different manifestations of the samedisease process. Patients with CD and IgAdeposits in their skin may represent ahitherto unrecognised subclinical form ofDH.

A specific small bowel enema appearance innon-responsive coeliac disease

N MIKiE, U UDESHI, J FERRANDO, AND P

ASQUITiI (East Birmingham Hospital,Birminghain) In untreated coeliac disease(CD) small bowel enema shows a reductionin the number of jejunal folds, and anincrease in the ileal folds, compared withnormal. A unique appearance, reversal ofthe normal 'ejunoileal fold pattern withatrophy ot the jejunum and 'jejunalisation'of the ileum is rarely found in longstandinguntreated CD.We have prospectively looked for this

reversal pattern. It was not found in 400patients with various non-coeliac gastro-intestinal diseases, nor in eight patients withCD who had a good response to a glutenfree diet (GFD). There was a significantreduction in the number of jejunal folds infive untreated CD and 10 non-responsiveCD patients, and a significant increase inileal folds (p<fl.02) in non-responsive CDwhen compared with all other groups, witha reverse pattern in nine of these 10.The 10 non-responsive CD patients had

persisting jejunal villous atrophy despite astrict GFD for over three years, and con-tinuing malabsorption. Five had co-existinghypersensitivity to foods other than gluten,four had coeliac lung disease, two hadcutaneous vasculitis, two associated coloniccarcinomas and one histiocytic lymphoma.Reversal of the jejunoileal fold patternspecifically identifies those coeliacs with anon-responsive state, who are likely tosuffer major complications of CD.

Bread is digested more slowly and incom-pletely if made from coarse flour

L J O)ONNELL, P M FMMETI, AND K W HEATON(University Department of Medicine, BristolRoyal Infirmary, Bristol) To discover ifflour particle size affects starch digestibilitywe gave 400 kcal breakfasts of wholemealbread and cheese in random order to ileo-stomates and normal subjects. Ileostomystarch excretion was measured for eighthours postprandially using amyloglucosi-dase, also blood glucose and plasma insulinfor three hours. With bread made from

A1352


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


coarsely milled flour, starch excretion(mean+SD) was 42% greater than withordinary fine flour (564±259 v 397+164,p=0-tl, n=9). At the same time, plasmaglucose and insulin responses were signific-antly lower. Thus coarse flour when com-pared to fine flour reduced the area underthe concentration curve for both glucose(90±46 v 154±82 mmol/l min, p<0-05,n=-9) and insulin (3407±1114 v 4138+1249mmol/l min, p<0-05, n=9). Swing ofplasma glucose from peak to nadir wasreduced with coarse flour (1-3+0-6 v1-8±0-6 mmol/l; p<0-(1) implying reducedinsulin action.Thus bread made from coarse flour

results in slower and less complete digestionso that more starch reaches the colon whereit adds to the substrate available forbacterial fermentation. This is consistentwith reports of increased faecal weight withcoarsely ground cereals. The glucose andinsulin data suggest that milling practicesdesigned to produce fine flour may havemetabolic implications for the consumer inrelaltion to obesity and diabetes.

Diarrhoea is related to fat malabsorptionand partial villous atrophy in HIV infectionin male homosexuals

A MlIl ER, P BATMAN, C FARQUHAR, S FORSTER,A PINChIING, W HARRIS, AND G E GRIFFIN (StGeorge's and St Mary's Hospital MedicalSchool, London) Jejunal biopsies weretaken from 20 HIV antibody positive malehomosexuals at different clinical stages ofHIV disease (asymptomatic six, persistentgeneralised lymphadenopathy nine, AIDSseven) after a 14C Triolein breath test(CTBT) to detect fat malabsorption. Eachbiopsy was microscopically examined formucosal enteropathogens and villus surfacearea to volume ratio (SV) was calculatedusing a Weibel graticule. Patients gradeddiarrhoeal symptoms as normal, mild (2-4),moderate (4-6) or severe (>6) based onstool frequency. A stool specimen wasexamined by culture and microscopy todetect enteropathogens.No enteropathogen was identified in stool

or jcjunal mucosa. Partial villous atrophy(PVA) was the sole histological abnorm-ality detected in 15 (75%) of biopsies andCTBT correlated with the degree of partialvillous atrophy (SV) (p<0.008). Thepresence of PVA and fat malabsorptionwere detected at all clinical stages of HIVinfection. Ten of 12 patients reportingmoderate or severe diarrhoea had evidenceof fat malabsorption whereas one of eight

patients with normal fat aibsorption hadsimilar diarrhoeal symptoms (p<l0() 1).Thus diarrhoea in HIV antibody positive

male homosexuals in the absence of anidentifiable enteropathogen is related tojejunal PVA and fat malabsorption. Suchdiarrhoeal disease can be found at allclinical stages of HIV antibody positiveinfection.

Sorbitol malabsorption in normal volunteersand in patients with coeliac disease

G R CORAZZA, A STROCCIII, R ROSSI, ANDG GASBARRINI (I Departmnent of MedicalPathology, S. Orsola University? Hospital,University of Bolognat, Bolognta, Italy)Sorbitol is a hexalhydroxy alcohol used acs asugar substitute in many dietetic foods andas a drug vehicle. Previous studies havesuggested that sorbitol ingestion may be anadditional cause of non specific gastro-intestinal distress. We evaluated sorbitolmalabsorption in 30 healthy volunteers,seven patients with untl-eated coeliac diseaseand nine patients with coeliac disease on agluten free diet by a four hour H. breathtest. After ingestion of test solutioncontaining 10 and 20 g sorbitol and of foursweets (6-9 g of sorbitol), 900o, 100%, and62% of healthy volunteers, respectively,showed a significant rise in H, excretionindicative of sorbitol malabsorption. Out ofall healthy subjects tested MOMOo after 20 gand 50%) after four sweets complained ofsymptoms of carbohydrate intoleranceduring the eight hour period after sorbitolingestion. At a 5 g dose given at concen-trations of2%,4%,8%, 16(M malabsorptionwas shown in 10%, 12°/, 22°/, and 43%'O ofthe healthy volunteers. Symptoms ofintolerance at 5 g were experienced only atconcentrations of 8% and 16%(. Unlikehealthy volunteers and coeliac patients on agluten free diet, 1())% of untreated coeliacsmalabsorbed a 2% solution of 5 g sorbitol.Our results show that sorbitol malabsorptionand intolerance may result from ingestion ofdoses and/or concentrations usually foundin many dietetic foods and drugs andunderline the need to consider sorbitolmalabsorption as a possible and hithertounderestimated cause of gastrointestinalcomplaints.

H2- or '4C-breath tests in the diagnosis ofsmall intestinal bacterial overgrowth?

M P O'CONNOR, M HEAIY, A KEIIEIY, C T

KEANE, R R O'MOORE, AND D[ WFIR (Dept.s of

('linical Medicine, ('linical Microbiologyanid Biocheminstry, .S't Jamsle<s' Hospital andi(lTrinitY College Dublin, IrCel(and(l) Whilst thegold standaird diaignostic test in small intest-inal bacterial overgrowth (SIB(O) remllainsjejunal aspirate and culture, poor paitielitacceptability limits the use ol this tech-nique. This study compares the results ofthree breath tests in the diagnosis of SIB()-xylose breath test (XBT). glycoelolate(GBT) and glucose hydrogen ((H 1T).

Controls were 17 volunteers with no evid-ence of Gl disease. Forty three pattientswere studied for possible SIBO with 26/48having an abnormal jejunal culture (> I()'org/ml). The three breatth tests were per-formed on eaich subject. `4('-breath testresults (XBT, GBT) were expressed as'")dose expired as 4CO./mmol C'OC and I .-breath results as arealulider I1 excretiotncurve (AH.AUC). Thc most sensitive para-meter for the GH,BT wats a 3 h A1 AU(JC(sensitivity 92(%O, specificity 79'),, I.he mlostsensitive parameter for both XBT and (,iG1Twas > 1 abnormal value over the test periodt(XBT-sensitivity 69(Yo, specificity 53((,,GBT-sensitivity 77Y,o, speeificity 53'Yo).Excluding gastroenterostomriy ((GF)pattients improved (GH,BT specificity to88(o as all seven GE pattients had abnormalGHlBT but only four haid SIBO(/H prodLuC-tion (threc hour AH,AUC) correlatcd withba.cterial count in the abnormal culturegroup (r=)-61. p<0c- I).

Although the GHL1BT requires furthcrmodification for adcquatc diagnostic use inGE patients, this study finds thatt only theGHrBT shows adequate sensitivity andspecificity for the diagnosis of SIBO.

Fasting breath hydrogen (H2) in coeliacdisease

G R CORAZ/A, A STRO()CIII, ANI) (G (GASBARRINI(I Departmenit (of Medical Pathology, S.Orsola UniversitY Hospital, UniversitY ofBologtna, Bologna, Italy) We studied thepossible clinical significance of the highbasal levels of H. by analysing the breathexcreted by the following fasting subjects:50 healthy volunteers, 149 subjects withfunctional bowel disorders, 16 patients withsmall bowel bacteriall overgrowth proven bybacteriology, 34 patients with untreatedcoeliac disease, 40 patients with coeliacdisease on a gluten free diet and 4(0 patientswith disorders of the small intestine otherthan coeliac disease (disease control). Thefasting levels of H. in untreated coeliacpattients (x22-5±+19 3 ppm) were signific-antly higher than those in healthy volun-

A1353


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroetnterology

teers (x558+3 1 ppm), patients with func-tional bowel disorders ( x6-6+4-4 ppm),treated coeliac patients (x9-9+8-1 ppm)and disease controls (x7-0±6-7 ppm). Nosignificant difference was found betweenpatients with untreated coeliac disease andbacterial overgrowth (5x 14 7± 14() ppm).The percentage of patients with raised H.fasting levels in untreated coeliac disease(58-8%) was significantly higher than that in

the other groups, except for the patientswith bacterial overgrowth (43-7%). In 14coeliac patients studied before and aftergluten-free diet, fasting H, levels decreasedfrom 26-6+18 to 11-6+ 1) ppm, becomingnormal only in those patients with recovery

of intestinal lesions. Our results show thathigh fasting H, levels are a frequent featureof untreated coeliac disease and that thereturn to normal of these levels is predictiveof villous regrowth.

Inhibition of starch absorption by dietaryfibre

O HAMBFRG, J J RUMESSEN, AND E GUDMAND-

HOYER (INTRODUCED BY P M CHRISIIANSEN)

(Gentofte University Hospital, Deptof Gastroenterology F, Copenhagen,Denmark) Using the H,-breath technic, theeffect of three dietary fibres on starchabsorption was examined in eight subjects.They received the test meals in randomisedorder: (a) bread made of 10) g wheat flour,(b) bread made of 10) g low gluten wheatflour, bread (a) with either (c) wheat bran,(d) beet-fibre or (e) pea fibre, and (f) thefibres seperately (15 g each). The quantityof H, excreted in excess of that produced bythe fibres alone was calculated.The percentage of malabsorbed starch

were: (a) 8%/ (4-17), (b) ()% (0-6),(c) 12,5% (5-22), (d) 12-5% (5-20), and(e) 12% (5-27). Low gluten bread was

malabsorbed, in small amounts, by onlythree subjects, p<0-05. The increase ofstarch reaching the colon when fibre was

added was: (c) 800% (: 14-133) p<0(05,(d) 36% (11-180) p<001 and (e) 660o(33-163) p<001. Orocaecal transit timeswere: (a) 360 min (210-420), (c) 165 min(120-300) p<001, (d) 240 min (90-360)p<0-05 and (e) 270 min (90-300) p<0 05.We conclude that the dietary fibres

impaired starch absorption, and this may bean important factor in the regulatory effectsof dietary fibres on bowel function. Thegluten moiety seems to possess inhibitoryqualities to starch absorption. All threefibres decreased orocaecal transit time.

Effects of vitamin E deficiency on smallintestinal transport function

M SATO, M (OSS-SAMPSON, D P R MULLIER, AND

P J MiLl A (Institute of Child Health, London)The gastrointestinal mucosa is vulnerable todamage by lipid peroxidation. Vitamin E(E) is a powerful membrane bound anti-oxidant present in high concentration in

enterocytes. We have studied the effects ofE deficiency on mucosal function. Twentydays old malc Wistar rats were placed on a

diet either deficient in E or one to which E(l10) mg/kg) had been added. Brush bordermembrane vesicles of small intestine were

prepared by a Ca " precipitation methodand glucose and sodium uptake measured.NatK+ATPase, cAMP and cGMP were

determined in mucosal homogenates. Na+stimulated glucose uptake was rapid with an

'overshoot' in both groups but after nine

months of E deficiency maximal glucoseuptake was significantly impaired (757± 116v 1152±324 pmol/h/mg protein; mean+ 1SD; n=5; p<0-05). H+ stimulated Nauptake was similarly reduced (380+±148 l743+ 181, p<0-02). In the absence of a Na+or H' gradient uptake of glucose and Na'was not affected by E deficiency. No signifi-cant differences in the Na+K'ATPaseactivity or cyclic nucleotide concentrationswere observed. These data show that intest-inal transport is disturbed in E deficiencyand suggest that this is because of an effecton apical membrane transporters ratherthan on passive permeability or the entero-cyte Na+ pump. We speculate that Edeficiency may contribute to the malabsorp-tion seen in some states of malnutrition.

Skeletal myopathy secondary to seleniumdeficiency

D A KELtLY, A SHiENKIN, A COL, AND J A

WALKER-SMITHI (Queen Elizabeth Hospitalfor Children, London, and Glasgow RoyalInfirmary, Glasgow) A two year old girlwith microvillous atrophy was maintainedon longterm home parenteral nutrition(HPN). She presented with severe myalgiaand regression of walking skills. Investiga-tions: normal full blood count, electrolytes,liver function tests and transaminiases.Serum concentrations of vitamins A, E, andB12, folate, zinc and copper were normal.Plasma selenium was very low at <0-05 mol/l(reference range (RR) 0 -2-() mol/l); RBCglutathione peroxidase was undetectable atl/g Hb and plasma glutathione peroxidasewas 6/1 (RR 90-35/1). RBC enzymes showednormal B,, B., B, status. Electromyogram

was mildly abnormal. Muscle biopsyshowed atrophy of type 11 fibres and muscleselenium was )-23 g/g dry weight (adult RR:0(6-0)97 g/g). Chest radiograph, ECG andechocardogram were normal. Fifteen gramsper day sodium selenite was added to HPNand no other changc was made. Withinthree days muscle pain had improved; shebegan to crawl again and by six weeks shewas walking normally. At three months.plasma selenium had risen to 0(53 [tmol/Iand RBC glutathione peroxidase was 15 It/gHb. Although biochemical sclenium deple-tion occurs in adults and children on longterm HPN, symptomatic deficiency is rareand poorly understood. The dramaticresponse to selenium repletion aloncsuggested that it caused the myopathy.Selenium is widely availlble in food butchildren with severe malabsorption on HiPNare particularly at risk from scleniutndeficiency and parenteral supplementationshould be routinely considered for thescchildren.

Intestinal pseudo-obstruction: is manometryhelpful?

D KUMAR, 1) 1. WINGATI, N S WIll.lIAMS, AND

H II -IIOMPSON (Surgical and GastroinitestinalScience Research Unils, The LonidoniHospital Medical College, Lonidoni) Thediagnosis of a primary disorder of smallbowel motility is difficult and frequentlymissed on clinical and radiological examina-tion. Manometric evaluation may help coii-firm the diagnosis and indicate a specificmotor abnormality so that managcmentmay be rationalised. We have carried outsmall bowel manometry in 14 patients (age19-60 years) with a suspected primary dis-order of gut motility and 30 control subjects(age 18-55 years). Small bowel manometrywas done using an eight port perfused tubeassembly and a low compliancc Arndorferpump. Manometric evaluation confirmedthe diagnosis in 12 of 14 patients (85'Yo), andthus eliminated a psychiatric disturbance.Motor abnormalities included bursts of con-tractions occurring every 20)-3t) minutes infive patients, abolition of motor complexesin one, tachygastria (5-6/min) in one(normal 3/min), fast duodenal contractions(16-18/min) in two patients (normal 9-11/min), retrograde peristalsis in one andabsence of motor quiescence following aPhase III front in three patients. None ofthe control subjects showed any of theseabnormalities. In the other two patients,based on normal manometry, it was pOss-ible to exclude a primary motility disorder.

A1354


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Intestinal manometry is thus an essentialinvestigation for the diagnosis and manage-ment of patients with suspected primary gutmotility disorders.

Ontogeny of small intestinal motor activity

W M BISSET, J B WArr, R P A RIVERS, AND

P J Mll.A (Department of Child Health,Institute of Child Health, WC], St Mary'sHospital Medical School, London) In thepreterm infant the characteristic patterns ofsmall intestinal motor activity are poorlydeveloped. Using continuously perfusedmultilumen nasojejunal catheters we havestudied the development of such motoractivity in 12 preterm infants (28-42 weeksgestation). Fasting activity was studiedlongitudinally on 28 occasions and on 23occasions the motor response to enteralfeeds was measured. With increasing gesta-tional age a clearly defined maturationof fasting motor activity was seen withincreases in: average duodenal pressure2-12 mm Hg (p<0-00)1); slow wave fre-quency 10-5-12-5 cpm (p<0.005); propaga-tion index 1-5 (p<0.001) and a progressionfrom random activity at 28 weeks, throughrepetitive fetal complexes to migratingmotor complexes by 37 weeks gestation.Nutritive sucking only occurred after 33weeks gestation. Bolus feeds appropriate tothe infants age produced no motor responsein the very preterm infants (28-30 weeks)but with increasing gestational age

(p<0)005), volume of feed (p<0-001) andtime exposed to feed (p<0001), the lengthof postprandial activity increased (0-140min). Disruption of fasting activity,however, did not occur when low volumecontinuous feeds were used. These datasuggest that the development of fastingmotor activity and nutritive sucking isgestationally dependent whereas the post-prandial motor response is dependent on

the pattern of feeding. We speculate thatthe former mirrors central and entericneural development whilst the latter is morelikely humorally related.

Effect of ileal infusion of fats, detergents andbile acids on stomach to caecum transit timein the rat

NICOLA J BROWN, A RICHARDSON, R D E

RUMSEY, AND N W READ (Sub-Department ofHuman Gastrointestinal Physiology andNutrition, The University, Sheffield) Wehave recently shown that ileal infusionintralipid (a triglyceride) significantly

increases stomach to caecum transit time(SCTT). This study aimed to investigate thenature of substances exerting this delayedtransit. Twenty two rats (250-300 g) withileal cannulae (20 cm proximal to theileocaecal valve) and six rats with isolatedileal Thiry-Vella loops (7 cm) cannulatedfor external infusion were used. After an

18 h starvation, test substances were infusedinto the ileum for 20 minutes (0-3 ml/h)before the rats were fed 5 ml beans bygavage. The infusion continued for a further160 min. Excreted hydrogen was monitoredcontinuously to give an index of SCTT.Delayed SCTT of the bean meal was

observed during infusion of a triglyceride(corn oil, p<002), a free fatty acid (oleicacid p<0.02), detergents (AOT, SLESp<0-02), a bile acid (deoxycholatep<0.05), a polyunsaturated oil (borage,p<0.05) and the fat analogue sucrose poly-ester (p<005) when compared with controlinfusions. Triglyceride infusion (intralipid)directly into isolated ileal loops also delayedSCTT (p<001), suggesting that breakdownof triglycerides to fatty acids is not essentialfor ileal inhibition of gastrointestinaltransit.

In summary we have shown that a rangeof lipid substances, detergents and bile acidswhen infused either directly into the intactileum or into an isolated ileal loop delaySCTF of a bean meal.

HLA D/DR Ag expression in the jejunum ofchildren with idiopathic protracted diarr-hoea (IPD) and circulating enterocyte anti-bodies (EcAb)

R MIRAKIAN, C A RICHARDSON, J A WALKER-

SMITH, S HILL, P J MILLA, AND G F BOTAZZO

(Middlesex, Queen Elizabeth Hosp5ital, andHospital For Sick Children, London) Smallintestinal villous enterocytes express ClassII molecules whereas crypt cells do not. Therole of such molecules in the gut is unknownbut it has been suggested that they may beinvolved in maintaining tolerance to oralantigens. In some children with IPD, circu-lating EcAbs are present together withother autoimmune manifestations. Jejunalbiopsies from nine children with IPD,EcAbs and an enteropathy were comparedwith histologically normal jejunum fromnine controls. We studied HLA Class I andII molecule expression in the enterocyteepithelium by immunofluorescence. HLAClass I products were normally expressed inthe jejunal epithelium of IPD patients.Class II (DR complex) expression in con-trols was restricted to upper villous entero-

cytes. In six of nine patients with IPD andEcAbs (titres > 1:64) aberrant cxpression ofDR molecules was seen in crypt entero-cytes. It can bc suggested that by analogywith a similar phenomenon detected inglands from paticnts with classical auto-immune diseases, DR +ve crypt enterocyteswould be able to act as Ag presenting cells,bypassing macrophage requirement. In theother three patients with EcAbs (titres<1:8) a decreased reactivity was seeni invillous entcrocytes. In all nine patientsnumerous DR +vc cells were present in thelamina propria compared to controls. Wepostulate that aberrant MCH moleculeexpression in the small intestine mightreflect loss of tolerance to luminal antigenwhich could. lead to autoimmune diseasc.

Portal and peripheral blood short chain fattyacid (SCFA) concentrations after caecallactulose installation at surgery

E W POMARE, S PETERS, ANt) A FISHIi R (Deparl-ment of Medicine, Wellinigtbon School ofMedicine, Wellington 2, New Zealand)The fate of the major SCFAs (acetate,propionate, butyrate). in man which resultmainly from the fcrmentation of carbo-hydrate is largely unknown. We have studiedfermentation in man by measuring thescSCFAs in portal and peripheral blood, afterthe caecal administraition of lactulosc topatients maintained on polysaccharide freefood before undcrgoing elective chole-cystectomy. At operation portal andperipheral venous blood samples were takenbefore, and at 15 minute intervals up to 60minutes following, caecal injection of either20 ml sterile saline (n= 12), or 67 g (n=6) orIt) g (n= 10) lactulose - samples wereanalysed by GLC. Fasting levels (n=28)were (,umol/l, mean+SD); portal acetate128±70-8, propionate 34-4+23-3, butyrate17-6± 18-5; peripheral acetate 77-2±56 1,propionate 37±+12, butyrate traces only.After lactulose there was a rapid rise inportal SCFA, with mean peak levels (tmol/I) after 10 g lactulose being acetate 197,propionate 277, and butyrate 17 3 - theresponse was dose related. No appreciablechanges in peripheral SCFA were found.

In conclusion, the caecal fermentation oflactulose is rapid in man with aippreciableincreases in portal SCFAs.

Gastrointestinal bleeding from other sourcesin patients with colonic angiodysplasia

A C STEGER, R B GALLIANt), A IIEMINGWAY, C B

WOOD, AND J SPE.NCFR (Depts of Suirgery atid

A 1 355


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Radiology, Royal Postgraduate MedicalSchool, Hammersmith Hospital, London)In 5% of patients with gastrointestinalhaemorrhage, the cause is obscure. Angio-dysplasia is increasingly found on colono-scopy or angiography to be the commonestcause of bleeding in this group. Angio-dysplasia may be present, however, but notactively bleeding.

In 71 patients with colonic angiodysplasiaand GI bleeding, 31 stopped spontaneously.Of 40 that came to laparotomy, 31 had aright hemicolectomy. In seven a smallbowel lesion alone was removed (Meckel'sdiverticulum four, ileal carcinoid one,jejunal A-V malformation one, duodenaldiverticulum one). In two patients, both aright hemicolectomy and excision of smallbowel lesions was performed.

In the right hemicolectomy group, threedied and seven rebled, four needing furthersurgery. In the small bowel lesion group,two of seven rebled, and one underwent aright hemicolectomy. Neither of the twopatients having both lesions removedrebled.

In 40 patients with angiodysplasia whocame to surgery, 16 (45%) had a secondlesion and four (10%) required furthersurgery. If the cause of GI haemorrhage isuncertain, full investigation includinglaparotomy is essential. If a synchronoussmall bowel lesion is found it should beremoved at the same time as the colectomy.

PAFDIATRIC SFCTION

Absence of secretory response in jejunalbiopsies from children with cystic fibrosis

C J 'I'AYI OR, P S BAXTER, J HIARDCASTLE, AND P T

HIARDCASTLE (Depts of Paediatrics andPhysiology, University of Sheffield,Sheffield) The demonstration of reducedchloride absorption and secretion acrosssweat duct and respiratory epithelia in cysticfibrosis (CF) suggests that the diseaseresults from a specific defect in chloridepermeability. Although GIT manifestationsof the disease are common and secretoryprocesses in the normal intestine exten-sively studied no attempts have been madeto identify a possible intestinal chloridetransport defect in CF.To address this problem the secretory

activity of jejunal biopsies from childrenwith cystic fibrosis has been investigatedusing a modified Ussing chamber tech-nique. Samples from three children with

cystic fibrosis failed to respond when chal-lenged with the intestinal secretagoguesacetylcholine and prostaglandin E,, whilecontrol tissues exhibited a rise in potentialdifference and short circuit current inresponse to these agents. Both groups oftissues generated an increased potentialdifference and short circuit current associ-ated with sodium linked glucose absorption.These results demonstrate for the first timethat the defect in chloride transportobserved in other epithelia in cystic fibrosisalso exists in the jejunum and could con-tribute to the intestinal effects of thedisease. The technique used should permitfurther studies of the basic defect and maybe of diagnostic value.

Incidence, morbidity and mortality ofCrohn's disease in Scottish children

ROGER BARTON, SANDRIA GILLON, AND ANNF

FERGUSON (Universitv of Edinburgh andWestern General Hospital, Crewe Road,Edinburgh) From Scottish hospitalinpatient records from 1968-1983, 282children were identified with the onset ofCrohn's disease (CD) at or before age 16.The incidence has increased four fold from6-6/million age <16 in 1968 to 26-4/millionin 1983. This increase occured within theages of 9-16, CD remaining rare under theage of nine. A representative sample of 68cases was investigated in detail.The mean age at onset was 12-6 years,

with a follow up of 7-05 (SD±4-2). Eachchild spent 81-5 days (range 7-322) inhospital, with a mean of 4-4 admissions up tothe age of 20 years alone. Disease distribu-tion and presenting features were similar topublished large adult series. Forty patients(72°/0) have already required surgery,14-5% (10) have a permanent stoma.During the period of the study there werefive deaths, a mortality rate of 7/0.

This unbiased, unselected cohort ofCrohn's patients show no unusual featuresof the disease itself when presenting inchildhood (although of course there aresignificant effects on growth and nutrition),however, a substantial social and clinicalmorbidity and mortality is seen.

Cow's milk sensitive enteropathy: animportant cause of failure to thrive in cysticfibrosis in childhood

S M HIll.l, M B MEARNS, AND J A WALKER-SMITH

(Academic Department of Child Health and

Cystic Fibrosis Unit, Qlueen ElizabethHospital for Children, London) Lactoseintolerance has been described in childrenpresenting with cystic fibrosis, but little ifany has been written about cow's milkprotein intolerance in this disease. Childrenwith cystic fibrosis commonly present withfailure to thrive. Causes include poordietary intake, recurrent chest infections,and pancreatic enzyme insufficicncy. Itfchildren attending a paediatric cystic fibro-sis clinic continued to gain weight at lessthan the normal velocity, even after ade-quate treatment for the above problems.small bowel biopsy was performed (sevencases). In five (aged less than one year) apatchy enteropathy compatible with a histo-logical diagnosis of cow's milk sensitiveenteropathy, was seen. In all cases removalof cow's milk (and lactose) from the dietresulted in improved weight gain, and stoolconsistency. Three underwent repeat smallbowel biopsy which was normal. Four toler-ated cow's milk by three years; the fifthremains on the exclusion diet at a year old.On review of the clinical features, all had

had cystic fibrosis diagnosed within a fewweeks of birth. Three of the five had pre-sented with meconium ileus (overall only10% do so).Cow's milk elimination has a significant

impact on failure to thrive in some patientswith cystic fibrosis (1:30 attending ourclinic). Presentation with meconium ileusappears to be common in this group.

Small bowel biopsy is an essential investi-gation in those with cystic fibrosis with poorgrowth in spite of adequate treatment.

Simplified and accurate oral pancreaticfunction test for use in children

J W L PUNTIS, J 1) BFRG, 1) SU F, AND I W BOOTil(Institute of Child Health, University ofBirmingham, Birmingham) The use of 14-Clabelled PABA in conjunction with the BT-PABA (Bentiromide) screening test forpancreatic insufficiency (P1) is unacceptablein children. We substituted p-aminlo-salicylic acid (PAS), a structural andpharmacokinetic analogue of PABA. for14-C PABA. This modified test was evalu-ated in 28 children with PI (25 cystic fibrosis,two Shwachman's syndrome, one pan-createctomy), aged 17 months- 16 years,together with a group of 20) control patientsaged 4 months-1 I years, shown to havenormal pancreatic function by pan-creozymin secretin testing. After an over-night fast, BT-PABA (15 mg/kg) with 4 5mg/kg PAS was administered with a stand-

A 1 356


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


ard breakfast of cereal and milk. Followingthis, a six hour urine collection was made.After the alkaline hydrolysis of PABA andPAS conjugates, urine samples wereassayed by HPLC. The PABA ExcretionIndex (PEI) was derived from the ratio ofPABA/PAS urine concentrations. Themean PEI in patients with P1 was 19 (range4-60) and for controls, 87 (range 66-140),with complete separation between patientswith PI and those without. We concludethat the exocrine pancreatic status of child-ren may be quickly and accurately determ-ined by a six hour oral test incorporatingBT-PABA and PAS.

Pathophysiology of childhood duodenalulcer

P K H TAM (INIRODUC[F) BY PROFESSOR

R SH1IEL1DS) (Department of Child Health,Royal Liverpool Children's Hospital, AlderHey, Eaton Road, Liverpool) Very littlepathophysiological data on childhood duo-denal ulcer exist. Is the childhood conditionsimilar to or different from the commonadult disease? We studied gastric acid secre-tion, emptying rate and serum gastrin con-centrations in a group of children withprimary chronic duodenal ulcer and a groupof normal children as control.

Both BAO and MAO were significantlyhigher in DU (n=30) (mean±SD=0 16±0 12 mmol/kg/h, 0-55±0+16 mmol/kg/h respectively) than in normal (n=33)(t)09±0*06 mmol/kg/h, 0-30±0-11 mmol/kg/h) (p<0-01). In addition patients withmore severe disease (n= 15) had higherMAO (0-63+0-17 mmol/kg/h) than thosewith mild disease (n= 15) (0-47+±017 mmol/kg/h) (p<0 02).

Gastric emptying rates were similar inDU (median=6-8 min) and normal (7-0min), (p>0-01). Fasting serum gastrin con-centrations were significantly higher in DU(n=25) (60-4+9-7 pg/ml) than in normal(n =25) (38 0+4-2 pg/mI) (p<0(05).Integrated gastrin response to meal stimula-tion was also significantly higher in DU(14-9+2-2 ng min/ml) than in normal(6-45±0-9 ng min/ml) (p<0.001). Therewas no significant correlation between acidsecretion and gastrin levels.Of the 25 DU patients, 11 had hyper-

gastrinaemia (mean±2 SD) alone, six hadacid hypersecretion alone, two had bothhypergastrinaemia and acid hypersecretionand only six had both normal gastrin andacid levels.Our study suggests that there are both

similarities and differences between child-

hood DU and adult DU. Unlike adults,paediatric patients not only have excessiveacid and gastrin response to stimulation,they also have increased basal acid andgastrin secretion. The prognostic value ofacid study in childhood DU is unique.Findings of gastrin study in childhood DU,never reported previously, suggest that thisgastrointestinal hormone is implicated inthe pathogenesis of the disease in more thanhalf of the patient population.

Campylobacter associated gastritis inchildren

M J MAlHONY, J WYATT, AND J M LI5FLEWOOD(INTRODUCED BY J KM.ILEHER) (Departmentsof Paediatrics and Pathology, St James'sUniversity Hospital, Leeds) Gastric coloni-sation by Campylobacter pyloridis isstrongly associated with non-auto immunegastritis in adults, where its significanceremains controversial. Its importance inchildren is unknown. Since 1981, 38 child-ren presenting with persistent epigastricpain (aged 1-16, mean 10(3 years, 21 boys)have undergone upper gastrointestinalendoscopy with gastric biopsy at thishospital. These biopsies were studied retro-spectively for C pyloridis colonisation usingthe modified Giemsa technique, and for thepresence of inflammation (Whitehead'sclassification). C' pyloridis colonisation wasdetected histologically in 9/38 (24%) cases,six of whom were boys. All patients wereaged 1t) or over, 9/28 (32%) being positivein this group. Eight of the ten patients withchronic gastritis were C pyloridis positive;one of six cases showing only small focalaccumulation of lymphocytes was positive,while all 21 with normal gastric histologywere negative. We detected C pyloridis in32% children aged 10 or over endoscopedfor epigastric pain. The association betweenCpyloridis and chronic gastritis in children,in whom gastritis is generally uncommon,supports a causal relation between thebacteria, gastric inflammation and epigastricsymptoms.

Peptic ulcer in childhood: the longtermprognosis

M S MURPHIY AND F. J EAS[HAM (Department ofChild Health, Royal Victoria Infirmary,Newcastle upon Tyne) Information islimited concerning the longterm prognosisfor paediatric peptic ulcer disease. We havefollowed up 19 individuals in whom thisdiagnosis had been made 14 to 27 years

earlier, based on strict diagnostic criteria.The subjects had ranged in age from 7 to 14(median 12) years at diagnosis and from 25to 38 (median 3(0) years at follow up. Nine(47%) had had a proven ulcer on investiga-tion, since entering adult life. Ten (53'.)were no longer prone to recurring episodesof abdominal pain but four of these hadundergone vagotomy and pyloroplasty forintractable symptoms in the past. Thus,only six patients (310%) appeared to havemade a lasting and spontaneous recovery.Complications such as haemorrhage,penetrating duodenal ulcer, severe pyloricstenosis or perforation had occurred in 10(53°%). Seven (37%) had undergonesurgery, and in two of these cases more thanone operation had been performed. Fiftyeight per cent of all complications sufferedand 89% of all surgical operations per-formed were in patients over the age of 21years. Our findings firmly reinforce the viewthat paediatric peptic ulcer disease fre-quently persists into adult life and a risk ofpotentially life threatening complicationsremains many years after diagnosis.

Gastro-oesophageal reflux in the preterminfant

S.l NFWFIL, I W BOOIHi, M E I MORGAN, ANI) A S

MCNEISH (Regionial Neonatal Intenisive CareUnit, Birmingham Maternity Hospital, anidInstitute of Child Health, University ofBirmingham, Birmingham) The incidence,severity and importance of gastro-oesophageal reflux (GOR) in the pre-terminfant is unknown. Twenty four hour intra-oesophageal pH monitoring was performedusing a novel 1 mm antimony pH electrode(Synectics Medical) and lower oesophagealsphincter pressure measured.

Fifty six measurements were made on 34patients (postconceptional age: median 31,range 26-29 weeks). Mean (±SEM) indicesof GOR were as follows: 12-1+2 episodes ofGOR per 24 h; pH was <4 for 45 ±0/ 9'Yoof the total time; the longest episode duringeach recording was 17±4-6 min. There wasno correlation between GOR and post-conceptional age, gestation, or lower oeso-phageal sphincter pressure.The effect of nursing care upon GOR was

assessed. Gastro-oesophageal reflux wasmost likely to occur at the time of physio-therapy, oropharyngeal suction, and nappychange (p<0(001), and was increased afterfeeds (0-05<p<() 1). Reflux was slightlyincreased in the left lateral position. Infantsreceiving xanthine for apnoea had a twofold increase in GOR (p<0)-05). A group of

A 1 357


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroenterolog,x

six infants with recurrent apnoea had mark-edly severe GOR (p<0.00(). A rapid reduc-tion in the frequency of apnoea followedabolition of GOR with thickened feeds.These results provide important physio-

logical data and have implications in thenutritional care of the very low birth weightbaby.

OESOPHAGUS

Factors influencing survival in oesophagealcancer

A WAI SON (Royal Lancaster Infirmary,Lancaster) Survival after resection foroesophageal cancer is poor, five year survi-val rate (5YSR) seldom exceeding 15% inthe West, although 5YSR greater than 50%)'for superficial lesions is reported fromJapan. It is unclear whether this relates toearlier detection or a different form of thedisease. This study was undertaken toinvestigate the relationship betweentumour staging and survival in 49 patientsundergoing resection between 1975- 81 pre-viously presented to this Society. Fortythree of these have now been followed for atleast five years and stratified according tocell type, depth of tumour invasion andcontiguous lymph node status.

Five year survival rate in resectedpatients was 14 3%, overall and 16.3% inthose surviving resection, representing29-4% for squamous and 7-7%/ for adeno-carcinoma. Nodal involvement occurred in58-8%Yo of squamous and 80-7% of adeno-carcinoma, where 5YSR was 10% and 0'/orespectively. In node negative cases, 5YSRwas 57.1% for squamous and 40% foradenocarcinoma. 23-6(o of squamous andno adenocarcinomas were confined to theoesophageal wall, 5YSR in this group being75%.

Depth of invasion and nodal status arebetter prognostic discriminants than celltype. Survival in the minority of lesionsdetected early is comparable to that ofJapanese series and to equivalent stages ofother tumours.

Laser therapy or intubation for palliation ofmalignant dysphagia

H BARR, N KRASNER, A RAOUF, AND R J WAL KER

(Gastrointestinal Unit, Walton Hospital,Liverpool) A prospective randomised trialwas done to compare laser therapy with

endoscopic intubation for inoperable oeso-phageal carcinoma. It was not possible todirectly randomise into laser versus intu-bation, as most patients were not suitablefor immediate intubation (because of com-plete obstruction or tortuous tumour).Patients were randomised into laserfollowed by repeat laser endoscopy everyfour weeks (L), or initial laser therapy toensure a suitable lumen to allow dilatationand intubation (1). Forty five consecutivepatients were treated (20)1 2(01 and fiveexclusions). Swallowing ability wasrecorded on a 0)-4 scale (0=swallowing allsolids, four=total dysphagia). Quality oflife was assessed by the QL index (scale1)-11)) and a linear analogue self assessment(LASA, scale 0)-20)0)). Mean swallowinggrade pretreatment was 2.9+±).8(L) and2-8+0-9(I) (NS). The best post treatmentwas 0.9+0.9(L) and 16±1)5(1) (p<().002).The QL and LASA improved by 017 and13-1(I) and by one and 20-1(L). Seven(I) and three (L) developed recurrentdysphagia with seven (I) and two (L) havingcomplications. Laser therapy offers moreeffective palliation for matlignant dysphagiathan intubation.

Manometric interpretation of the oeso-phageal egg transit test: a useful screeningtest for oesophageal motility disorders

C A ERIKSFN, R J 010LDSWORti, D SUTITON,N KENNFDY, AND A CUSCHIERI (Departmentsof'Surgery and Nuclear Medicine, NinewellsHospital & Medical School, Dundee) Theuse of manometry to detect oesophagealmotility disorders requires expertise in itsperformance and interpretation. The oeso-phageal egg transit test (OET) uses astandardised 99mTc solid egg bolus toobjectively and physiologically evaluateoesophageal transit. We assessed theoesophageal motility of 102 symptomaticpatients using oesophageal manometry andOET. Of 32 patients with normal OET, 10(310%) had abnormal manometry (7/11)being diagnosed non-specific oesophagealmotility disorder). Of 70 patients withabnormal OET, 51 (730%) had abnormalmanometry (x2=1582, p<0-001). Thecomputer generated condensed imagedefined five different transit patterns:normal (n=32); oscillatory (n=21); non-clearance (n= 16); 'step' delay (n= 16);miscellaneous delay (n= 17). Normal OETcorrelated significantly with normalmanometric parameters. OscillatoryOET showed significantly more tertiarycontractions, low amplitude waves and non-

propagating swallows, and confirnied themanometric diagnosis of achalasia (6/6) anddiffuse oesophageal spasm (2/4). Theremaining three patterns of abnormal OE ldid not significantly correlate with specificmanometric parameters or symptonil score.The solid bolus oesophageal transit testprovides an objective screening test of oeso-phageal motility disorders and should beperformed before oesophageal manometry.

H2 antagonists in oesophagitis: can physio-logical studies predict the response?

I) A F ROBTIRSON, M A ALDERSIFY, 11

SHEPIIPRI), R S lLOYD, AND C 1. SMI-I1I1 ()('-partment of' Medicinie I1, SouthainptonGeneral Hospital, Southampton) Theresponse to H. antagonists in reflux oeso-phagitis is disappointing with only 50)%/of patients healing. In an attempt toidentify factors predicting response wehave measured; ambulatory oesophagealpH, oesophageal manometry and fastingserum gastrin concentrations in 28patients with reflux oesophagitis, beforeand during treatment with ranitidine31)1) mg bd. Fourteen patients healedendoscopically at six weeks (group A)and 14 had residual oesophagitis (groupB). Group A were characterised by alower serum gastrin level before treat-ment (4-52 pmol/l; 24-10): mean andrange) and group B (11*1 pmol/l; 3-5-21:p<0-05) and showed a marked reductionin acid reflux on treatment to nearnormal values. Mean /O time below pH4fell from 14-9 to 4-2 in group A (p<0-05)but was not affected in group B ( 14-2-15-6, not significant). Severity of reflux.degree of oesophagitis, lower oeso-phageal sphincter tone (Group A 27-8 mmHg, group B 26-8 mm Hg), age, sex,smoking, obesity, duration or severity ofsymptoms were not significantly associ-ated with healing. Abnormal peristalsiswas common (29 4% abnormal wavesgroup A, 36% group B, NS) but did notinhibit the response to ranitidine, anddid not improve with healing. Healingwith H. antagonists in oesophagitis isdue to reduction in acid reflux, with noeffect on oesophageal motility. Theresponse cannot be predicted beforetreatment.

Treatment of reflux oesophagitis with aprostaglandin analogue

HI L SMART, P D JAMES, M AFKINSON, AND C J

IIAWKEY (University Hospital, Nottingham)

A 1 358


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


The prostaglandin analogue trimoprostil isgastro protective, inhibits acid secretionand does not reduce lower oesophagealsphincter tone. We therefore assessed itsvalue in reflux oesophagitis.Twenty two patients received trimopro-

stil 750 btg qds and 22 received placebo in aone month double blind controlled clinicaltrial. Both groups were allowed antacids.Patients treated with trimoprostil (but notplacebo) experienced a significant reduc-tion in the frequency of pain, from seven(median, range 1-7) days/week (p<0.005)and in its duration, from two (1-12) to 0-5(0-6-5) hours/day (p<0-02). Both groupsreported a reduction in the severity of painand sleep disturbance during the trial butdifferences between them did not reachstatistical significance. There was a sig-nificant reduction in the extent of oeso-phagitis seen at endoscopy after treatmentwith trimoprostil from 5 (1-9) cm to 3 (0-9)cm (p<0(00l). At the end of the trial eightpatients taking trimoprostil showed noendoscopic evidence of oesophagitis.

Trimoprostil was well tolerated andappears to bc an effective treatment foroesophagitis. Prostaglandins may be cyto-protective to the human oesophagus.

Ultrasound identification of oesophagealvarices - comparison with endoscopy

S H SAVERYMUITIU, A E A JOSEPH, AND J D

MAXWI.I I. (Depcartments of Medicine II andUltrasoutd, St George's Hospital andMedical School, London) Oesophagealvarices are an important feature of chronicliver disease and are currently recognised byendoscopy or radiology. Because ultra-sound is routinely used as the initial assess-ment of chronic liver disease the identifica-tion of oesophageal varices at this timewould aid both investigation and manage-ment. We have attempted to directly visua-lise oesophageal varices by ultrasoundexamination of the lower oesophagus in aprospective comparative study with uppergastrointestinal endoscopy in 100 patientswith suspected or proven liver or biliarytract disease. Oesophageal varices wererecognised on ultrasound by thickening ofthe oesophageal wall and irregularity of theair containing lumen. In seven patientsultrasound was unable to adequately visua-lise the oesophagus while in four patientsendoscopy was equivocal. Excluding thesetwo groups ultrasound detected 18 of 22oesophageal varices (sensitivity 82%)including all patients with medium or large

varices while incorrectly suggesting varicesin their absence endoscopically in sixpatients specificity 910.We conclude that ultrasound can identify

oesophageal varices with high sensitivityand specificity and should be specificallyexamined for in all patients with suspectedchronic liver disease.

Prospective controlled trial of propranololand sclerotherapy for prevention of rebleed-ing from oesophageal varices

C VICKI-RS, J RHOI)ES, P HI111ENBRAND,H BRADBY, P HAWKER, P D)YKIS, I CII[SNER,R COCKEL, D ADAMS, R VAl ORI, J DAWSON,H O'CONNOR, AND E ELIAS (Qtueeni Elizabeth,Good Hope, Genieral, Sell*vOak, EastBirmingham, acnd Santdwell Hospitals,Birmingham) Sclerotherapy is of value inthe treatment of acute variceal haemorr-hage although rebleeding occurs in up to55% of patients. We have evaluated the roleof additional longterm propranolol tofurther reduce the rebleeding rate andimprove survival. Sixty nine patients (41M), admitted with bleeding varices to adistrict hospital, were randomised toreceive regular sclerotherapy, plus pro-pranolol 160 mg (long acting) (n-35) orplacebo (n=34). Median follow up was 96and 110 weeks respectively (range 44-233).Both groups were similar for age, sex,aetiology, and Child's grading.

Interim analysis on 'intention to treat',55% of propranolol and 71% of placebotreated patients were free of rebleeding(p=NS). Median time to rebleeding was 15weeks (1-130) in the propranolol and nineweeks (1-94) in the placebo group. Therewere 22 deaths (propranolol 13) withrebleeding responsible in 13 (propranololnine). Cumulative per cent survival was58% in the propranolol and 72%'/ in theplacebo group (p=NS).The results show that in patients under-

going regular sclerotherapy after an episodeof variceal haemorrhage, longterm pro-pranolol confers no additional benefit insurvival or prevention of variceal rebleed-ing.

GAL LBLADDER

The British/Belgian Gallstone Study Group's(BBGSG) postdissolution trial

K HOOD, D GlIEESON, D C RUPPIN, R I1DOWLING, AND T1HE BBGSCi (Gastroenterology

Uiit,homa'

s Campus, UMDS of GuY s & StThomcas'Hospitals, Lonidoni) The BBGSG'strial was a prospective, multicentre, randomallocation trial of low dose ursodeoxycholicacid (UDCA) 3 mg/kg/day, placebo (doubleblind) or high fibre low refined CHO diet inpreventing gall stone (GS) recurrence inpatients with confirmed complete dis-solution. Ninety three patients, stratifiedfor obesity and stone free intervail sincedissolution, have been followed for up tofive years (28+SEM 1-5 mo) with six monthsultrasound (US) and yearly cholecysto-grams (OCG).There were 21 recurrences, only two

symptomatic, 21±2 mo (range 12-42) aftertrial entry, equivalent to 9'0±3-2'2, at 1,23-4±500 at 2, 30-5+6%/, at 3 and 35-5±7-5at three and a half and subsequent year, bylife table analysis. Recurrence was greaterand earlier by US than by OCG. There werefour recurrences in the UDCA, six in theplacebo and 11 in the diet groups, corres-ponding to 23-5±10(8/%, 30-6±118Xo and46-4 14-2% respectively at three and a halfyears. Recurrence rates for UDCA and diettreated groups were not significantly differ-ent from those taking placebo. Age,obesity, weight change and pregnancy didnot affect outcome, but men showed morefrequent recurrence than women (NS). Thestone-free interval between dissolution andtrial (median 9 mo), was important - therecurrence rate at three and a half yeairbeing 62-9±13-4% for <9 mo and 15 7+7-50/, for >9 mo (p<0-02).

Neither low dose UDCA nor diet areeffective in preventing GS recurrence.Patients who remain GS free for several mowithout treatment constitute a 'low-risk'group. More than 50% of patients remainGS free for up to five years.

Interdigestive gall bladder emptying and itsrelationship to motor activity in man

S ELI-ENBOGEN, J S GRIME, J CALAM, C RMACKIE, S A JENKINS, AND J N BAXTI R(Department of Surgery, University oJLiverpool, Department of Nucear Medicine,Royal Liverpool Hospital, Dept of Gastro-enterology, Royal P G Medical SchoolLondon) Human gall bladder emptying(GBE) occurs during both digestive andinterdigestive periods. The mechanism forinterdigestive GBE is unknown but may berelated to the intestinal migrating motorcomplex (IMMC) (fasting cyclic and phasicintensity activity). Using "Tc'0-EHIDAcholescintigraphy and gastrointestinillmanometry, interdigestive GBE and the

A 1 359)


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


IMMC were observed in two groups ofhealthy volunteers; group I, observed for134-298 minutes without stimulation (n=18); group II, observed for 90) minutes afteratropine (0-6 mg+0'3 mg/kg/h) (n=8). Ingroup 1, 17 complete IMMC cycles were

recorded. The cycle length was 100±15-5minutes, and the lengths of the phases were;

phase I, 35-3±6-6 minutes; phase II,

60(6±10 4 minutes; and phase III, 5-2±0 3minutes. Interdigestive GBE occurred on

nine occasions; four during phase I, and fiveduring phase II. In 10 of the IMMC cycles,GBE did not occur, whilst in two cyclesGBE occurred twice. The relationshipbetween spontaneous GBE and the phasesof the IMMC was random (p<025,X2goodness of fit). The gall bladder emptied181±0-4% of its contents (mean±SEM).From group I, the probability ofGBE in anyminute period was calculated as 0-0053.Atropine inhibited GBE in all eight group II

volunteers during the 90 minute studyperiod (p<0001, Poisson test, group II v

group I).These results suggest that interdigestive

GBE (1) occurs infrequently and in smallvolume, (2) is unrelated to the phases of theIMMC, (3) is cholinergically mediated.

Bile lipid and gall stone (GS) dissolutionresponses to combined chenodeoxycholicacid (CDCA) and ursodeoxycholic acid(UDCA) treatment

K HOOD, D GLEESON, G M MURPHY AND R H

DOWlINCi (Gastroenterology Unit, Guy'sCampus, UMDS of Guy's and St Thomas'Hospitals, London) Oral CDCA dissolvesradiolucent GS with moderate efficacy butcauses dose related diarrhoea and hyper-transaminasemia. Ursodeoxycholic acid isfree from these side effects but has lowerefficacy, in part because of acquired GScalcification. To minimise side effects whilstmaintaining efficacy, we gave 7 5 mgCDCA plus 5 mg UDCA kg/day (halfnormal doses) to 34 patients (6 M, 28 F)aged 23-91 (median 51) yr with radiolucentGS measuring 8 7+SEM 0 9 mm max diamin 'functioning' gall bladders. Biliary cholsatn index (SI), measured in 15, fell from1 11±0 09 before (SI<1 0 in five suggestingnon-chol stones) to 0(8±0)06 after six weekstreatment. Fourteen were treated <6 mo

(surgery in four, default in three, three tofive months treatment in seven). Of 20followed >6 mo, three showed completeand five partial GS dissolution correspond-ing to 44-4±12 5% at one year by life tableanalysis. Two patients showed no response

at six months but partial GS dissolution at12 months. No patients with suspected non-chol stones responded to treatment. Therewere no abnormalities of liver function orserum lipids and no acquired GS calcifica-tion but 11 had diarrhoea- transient in four,but requiring CDCA dose reduction in fiveand withdrawal in two.

(1) Pretreatmcnt SI <1 0 predicts lack ofresponse. (2) CDCA+UDCA have lowerefficacy (44-5±12 5% at one year) but fewside effects than single bile acids. (3) With-drawal of treatment for non-responseshould be delayed until one year.

In vivo choledocholitholysis using methyltertiary butyl ether (MTBE)

G A LAFERILA, G FULLIERTON, AND W R MURRAY

(University Department ofSurgery, WesternInfirmary, Glasgow) The management ofcholedocholithiasis has been revolutionisedby endoscopic sphincterotomy. Stoneextraction through the sphincterotomy islimited by their size. If extraction fails anasobiliary catheter (NBC) can be placedendoscopically thus allowing access to thestones for dissolution therapy. We presentour preliminary experience using thealiphatic ether MTBE. Ten patients (eightmen, two women; age range 65-85 years)with suspected choledocholithiasis werereferred for ERC, six having previouslyundergone cholecystectomy. Endoscopiccholangiography confirmed multiple (2-12)large stones (diameter 15-I mm-27 5 mm).Post sphincterotomy extraction failed in allcases. An NBC was inserted and 24 hourslater MTBE dissolution was commencedusing 2-5 ml of the solution every 3t)minutes over a two hour period. An averageof four dissolution sessions were used perpatient. Subsequent stone assessmentrevealed complete disappearance in threepatients, significant reduction in stone sizeallowing subsequent duct clearance in fivepatients and no change in two. Complica-tions noted were mild drowsiness (onepatient) and nausea (one patient). Minortransient abnormalities in liver functiontests occurred in all patients. This effectiveagent, which has a low morbidity may be ofvalue in the management of the frail andelderly patients with large bile duct stones.

Gall bladder motility in the presence of gallstones

G T SUNDERLANI), C Gi SUIHERILAND, AND D C

CARTER (University Departments of Surgery

and Pathology, Royal Infirmary, Glasgow)Both increased and decreased gall bladdercontraction are described in the presence ofgall stones and an altered sensitivity tocholecystokinin has been implicated. Wehave examined in vitro gall bladder con-tractility and sensitivity to cholecystokininoctapeptide (CCK-8) in 42 patients with gallstones and in 14 'control' patients havingcholecystectomy during hepatic artery can-nulation. Full thickness strips from the bodyof each gall bladder were suspended inoxygenated Krebs' solution at 370C. Doseresponse curves were obtained to CCK-8(10 "-10( M). The concentration ofCCK-8 required to achieve 50% of maxi-mum contraction (ED,,,) was taken as ameasure of sensitivity. Muscle contractilitywas derived from the maximum contractionand the muscle volume of each strip.Median contractility in the normal groupwas 60 g/cm (range 28-96) which wtassignificantly different from gall bladderswith stones (median 25 g/cm3, range 4-1t)2)(p=t)0002 Mann Whitney). There was nodifference in ED,, values between thegroups (normal: median 40-7 nM, range0-05-100) (gall stones: median 40(7 nM,range 9-200).We have confirmed that gall stoncs are

associated with a significant reduction ingall bladder contractility without anychange in sensitivity to CCK-8.

Stone extraction after endoscopic sphinc-terotomy - and active policy is best

D F MARTIN, J C MCGREGOR, M E LAMBERT, ANDD E F [WEEDLE (Gastrointestinal Unit,University Hospital of South Manchester,West Didsbury, Manchester) In order toevaluate a policy of primary stone extrac-tion after endoscopic sphincterotomy (ES)we have reviewed the success and complica-tions of this procedure in 262 patients seenbetween January 1981 and December 1985.In 85 patients whose bile ducts were clearedof stones immediately after ES, acute pan-creatitis developed in three (3.50/o) andanother (1 22%) died from a pancreaticabscess after initial admission with acutepancreatitis. None developed any othercomplication. In 177 patients whose stoneswere not cleared, 56 had insertion of apernasal catheter. Eight of these (10(2%)developed complications, two (3.4%o) ofwhich were cholangitis related to tubeblockage or displacement. In the remaining121 patients, stones were left without drain-age. Complications developed in 18(15.2%) but eight of these had bleeding

A1360


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


which precluded attempts at extraction.Four patients (3-3%) developed cholangitisalnd another (0(8X%) cholecystitis despiteintravenous antibiotics. Two patients(I *6%) died. Although development ofcomplications of ES, particularly bleeding,may lead to failed duct clearance, it appearsthat infective complications are more likelywhen duct clearance fails. An active policyof primary stone extraction with the use ofpernasal catheters or stents when this fails,may be expected to reduce the risk ofcholangitis.

C(01ORECTAI I POSTI,RS

Can videoproctography and anorectalphysiology predict outcome after rectopexyfor the solitary rectal ulcer syndrome(SRUS)?

I (, FINIAY, C' I BARIRAM, AND R J NICHOLLS(St Mark's Hosspital, City Road, London)Anterioposterior rectopexy improves somepiatients with SRUS without external pro-lapse. We have carried out preoperativcvideoproctography and anorectal physio-logy in 17 such patients in an attempt topredict the outcome of surgery. Postopera-tively 12 (group A) were improved sympto-matically and five (group B) were not.Videoproctography showed complete rectalemptying within 30 seconds in nine patientsand incomplete emptying in three patientsin group A. All five patients in group B hadincomplete emptying (p<0t)5). Perinealdescent, estimated in 14 patients, was foundin 4/9 in group A and 5/5 in group B(p<0-05). Internal prolapse was found in6/12 patients in group A and 5/5 in group B(p<t)08, NS). Eleven of the 13 patientswithout paradox were in group A.

'[here was no significant differencebetween basal/squeeze anal canal pressures(group A 96±14 mm Hg/65±16 mm Hg,group B 77±18 mm Hg/70±10 mm Hg).Pudendal nerve latencies in groups A and Bwere 2-16+0-15 ms and 283+0±0(34 ms(p<0()8, NS).

Incomplete rectal emptying and perinealdescent on videoproctography appear topredict a poor outcome after antero-posterior rectopexy for SRUS.

Consumer guide to anal manometry: a com-parison of microtransducer, water and airfilled microballoon systems

R MILLER, D C C BARTOtO, A M ROE, N J MCC

MORTENSEN, AND [) JAMES (Department ofSurgery, Bristol Royal Infirmnary, Bristoland Department of Clinical Measurement,Royal Devon and Exeter Hospital, Exeter,Devon) Water filled microballoon mano-

metry systems have many disadvantagesand microtransducer tipped catheters(MTC) have been advocated as reliablealternatives. Maximum resting (MRP) andmaximum voluntary contraction (MVC)pressures were compared using the twosystems with a standard station pull throughtechnique in 12 patients. There was poor

correlation (MRP r,=0-62 p<0(05, MVCr,=0-42 p>0-05). To determine whetherthis was because of MTC radial pressurevariation we studied 39 patients with bothsystems, recording pressures from each offour quadrants with the MTC. Improvedcorrelation was found. (MRP r, 072,MVC r=0-087 p<)-001). Repeated MTCmeasurements was a laborious process andwe therefore compared a new air filledmicroballoon system with the water filledsystem in 44 patients with an excellentcorrelation (MRP rs=)-86 p<0()-0l andMVC r,=094 p<0-)01). Repeat studies in1i) patients showed good reproducability(sphincter length r,= 095, MRP r,=0-98,MVC r,=-089).These results show the limitations of

microtransducer catheters. Air filled micro-balloon manometry correlates well with a

water filled system and has many advant-ages over both the alternatives studied.(r, Rank Spearman correlation coefficient.)

Anorectal sensation: its role in the contin-ence mechanism

R MILLER, D C C BAR'101O, F CERVIRO, ANI) N J

MCC MORIENSEN (Dept of Surgery, BristolRoyal Infirmary, Bristol and Dept ofPhysiology, University of Bristol, Bristol)To elucidate the role of sensation in thecontinence mechanism we studied mucosalelectrosensitivity (ME) and temperaturesensation in the lower, mid and upper zones

of the anal canal in 40 control patients (age43, 18-83)*, 24 with incontinence (age 57,34-8t))* and 20 with haemorrhoids (age 55,23-77)*. A small electric current passedbetween two electrodes 1 cm apart on a

catheter measured ME in millivolts and a

water perfused thermode I cm long was

used to measure the minimum detectabletemperature change (MDTC) in 'C. Theincontinent group had a severe sensorydeficit in all zones and the haemorrhoidgroup in general had a milder sensory loss

(control/haemorrhoid/incontinent; ME(mv): lower 4-0/4-3/7-0', middle 4t)/5 55'8X0", upper 5-5/12-1'/1-5", MDTC °C:lower 08-/1-2'/2-2", middle )-9/1-2|/2-22,upper 1(0/3- 1"/>4-5').We believe that impaired anal sensaition

may be an important contributary factor tothe pathogenesis of faeccal incontinence.Lesser degrees of sensory loss related tohaemorrhoidal prolapsc is not associatedwith incontinence in the presence of goodsphincter function.

(Significant difference from controls:a=p<t)05, b=p<0)-0)01, Mann-Whitney Utest) * =median (range)

The sampling reflex: a comparison of normaland incontinent patients

R MILIER, D C C BAR'I'OLO, F CEIRVERO, ANt) N J

MCC MORiTENSEN (Departmlent of SurgerY,Bristol Royal Infirmary, Bristol an(l Depart-ment of Physiology, University of Bristol,Bristol) To investigate the concept ofanorectal sampling we studied 18 patientswith faecal incontinence and 18 age and sexmatched controls. A multichannel micro-transducer catheter was positioned so thatpressures were recorded from the rectumand the junction of the mid and upper thirdsof the anal canal. Resting pressures wererecorded for five minutes, and whilst dis-tending the rectum with It) ml increments ofair injected into a small latex balloon, andthen freely into the rectum. Sampling,(equalisation of rectal and upper analpressures), occurred spontaneously in 16 ofthe controls and only six of the incontinentgroup (p<t)-02, X2 test). Induced samplingoccurred at a higher rectal volume in theincontinent group for freely injected air(control: 10 ml lit)-70) mll, incontinent 40)ml 11[) ml->10() ml]- p<t)002, Mann-Whitney U test). There was no differencefor air injected into the balloon (control 20)ml [10-70 ml], incontinent 20) ml 110-70ml].We believe this defective sampling

response may be a significant contributoryfactor in the pathogenesis of faecal incon-tinence, preventing sensory discriminationof rectal contents. (a=median, range.)

Defecographic findings in young healthyvolunteers

P J SHORVON, F MCIIUGII, S SOMERS, AND G W

S'I EVENSON (Department of Radiology,McMaster University, Hamilton, Canadaand Department of Medicine, Toronto

A 1 36 1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society ofGstroenterology

University, Canada) As previous reports ondefecography suffer from the absence ofdatita on normal subjects, we recruited 48healthy young adult student volunteers (25men and 23 women) to undergo a defeco-gram. For the procedure, after coating ofthe rectal mucosa with liquid barium, abarium paste was inserted into the rectum.The external anal orifice was marked withbarium ointment, and women inserted acontrast soaked vaginal tampon. Videofluroscopic recordings in the lateral positionwith 10() mm spot films were taken at rest,with contraction of pelvic muscles, onstraining and during defecation. Twentyvariables were measured directly using acentimetre midline scale. Ten radiographicappearances were also recorded andgraded.

T'he results indicate the range of norm-ality to be much wider than previouslyrealised; for example 44% of men and 45%0of women had a high grade intusussception,77% of women had an anterior rectocoele(moderate to large in 9/23). The meanpelvic descent from rest in women was 2-0cm and greater than 3 cm in 23%. In menthe mean was 1.9 cm, with 20()/ greater than3 cm. Four subjects had an open anal canalat rest with apparent incontinence. Cautionmust be exercised in overinterpretingdefecograms of patients.

Physiology of defecation - the importance ofthe levator ani muscles

I (G FINILAY, K (CAR'I'ER, AND) I MCILOED (INTIRO-D)UCIH) BY D C CARTIFR) (University DeptSlurgery, Royal Infirmary, Glasgow)Coordinated relaxation of the puborectalis/anal sphincter muscles resulting in anobtuse anorectal angle is necessary forunobstructed defecation. The factors whichinitiate this mechanism are unknown. Weused videoproctography. micropressuretransducers and EMG to simultaneouslystudy anorecal angle, intrarectal/anal canalpressures and external sphincter/puborectalis/levator EMG in five normalpatients. The call to stool was initiated bythe instillation of 180 ml barium (1-25mg/ml) over a 120 second period. Observa-tions were made during the insitillation aindvoluntary expulsion of barium.The instillation of barium produced (a) a

mean fall in anal canal pressure from 40± 16mm Hg (mean±SE) to 13±8 mm Hg. (b) Arise in intrarectal pressure from 0 to30(5+12 mm Hg (c) An increase inanorectal angle from 86±5 degrees to103±7 degrees at the call to stool and

115+0-5 at 120 seconds. EMG showed nochange in puborectalis/anal sphincteractivity during the instillation of barium.Expulsion of barium was achieved byfurther increasing anorectal angle to 134±5degrees and decreasing anal canal pressureto zero mm Hg as a consequence of aboli-tion of puborectalis/external sphincterEMG activity. These results suggest thatelevation of the levator ani occurs beforerelaxation of the puborectalis duringdefecation. Failure of this coordinatedactivity may be the basic abnormality inpatients with symptoms of obstructeddefecation.

Does the position of the patient influence theresults of anorectal manometry?

K YOSHIOKA, V POXON, AND M R B KEIGHILI Y

(Surgical Department, The GeneralHospital, Birmingham) We have comparedmeasurement of pressures at 2, 4, and 26 cmfrom the anal verge using an open tippedperfused catheter in three groups ofpatients: controls (n= 1), constipation(n=21) and incontinence (n= 14). Pressureswere measured at rest (R), during maxi-mum pelvic floor contraction (Sq) andattempted defecation (St) in the left lateral(LL) and seated position (S). No significantdifference between the seated and lateralposition was observed at 2 cm in all groups.At 4 cm, there was no significant differ-ence in control group, whereas significantdifferences were observed in the group ofpatients with incontinence (R:p<0(025,Sq:p<0).0)25) and the group of patients withconstipation (R:p<0(025, St:p<0.()5). At26 cm pressures were significantly greater inthe seated position for patients with incon-tinence (R:p<0.01, Sq:p<)-0 1). Compli-ance (ml/cm H20) did not differ significantlywith position in controls (LI-848,S= 13-08), incontinence (LL= 14-82,S=16-34) or constipation (LL= 10-97,S=11-77). Measurement in the left lateralposition is unphysiological and differs sig-nificantly from the seated position particu-larly in incontinent patients.

Physiological parameter should dictate thesurgical management of longstanding idio-pathic chronic constipation

K YOSIIIOKA AND M R B KEIGHLEY (SurgicalDepartment, The General Hospital,Birmingham) We have found that idio-pathic longstanding constipation may be

due to outlet obstruction (00) as shown byfailure to evacuate rectal contents onproctography or by electromyographicevidence of increased puborcctalis activityon attempted detecation (n =26). Alternat-ively, patients may have colonic inertia (CI)as demonstrated by failure to achieve anincreased motility index after rectalbisacodyl or where passage of radio opaquemarkers is delayed (n=4). Some paitientshave both 00 and Cl (n= 10). Anorectalmyectomy (ARM) gives good results inpiatients with 00 whereas subtotalcolectomy is reservcd for patients with Cl.We hiave reviewed the results of ARM for00 with normial transit (60% 100%marker passed in five datys) and those with00 and/or CI (less thian 60%'/ markerspassed five days). ARM aIchieved passageof at least three stools per week withoutlaxative in 1t) of 12 patients (83%O) with 00who have normtil transit compared withonly eight of 16 patients (500/%) with 00and/or Cl. Six patients with 00 and/orCI have subsequently required subtotalcolectomy which restored bowel habit tonormal in five paitients (83%O) and nonedeveloped incontinence after a previousARM.

Clinical and physiological evaluation ofpostanal repair

K YOSHIOKA, (G IIYILAND, ANI) M R B K(GilIl I-EY(Surgic al Department, Th)e GenierailHospital, Birmin1gham) Postanal repairl(PAR) is used for treatmenit of idiopathicfaecal incontinence (IFI) on the premise thatit restores a deficient anorectal angle. Wehave rcviewed the longterm results of PARin 124 patients between 1976 and 1986. Wchave measured anal and rectal pressurcs,rectal compliancc, anal and rectal sensattionand performed proctograms before andthree months after operation (n=9).Results arc compared with age and sexmatched controls. Only 24 (34%O) of the 79patients followed for more than three yearsclaimed complete continence. Although 50(69%/) were improved, soiling occurred in43 (62%O) and 4t) (56(%) continued to wearpads. Resting (R) aind squeeze (S) analcanal pressurcs were significanitly lower inIFI thanl controls (R:49-7 v 83-0, p<0-0l,S:77-2 lo 130(9, p<0-005) there 'was nosignificiant improvement in anal pressuresafter PAR (R before and after: 49-7 v 47-3,S before and aifter: 77-2 v 691, NS).Furthermore there was no objectiveimprovement in anal sensation (lower zonebefore and after PAR: 137 mA and 144 mA,

A1362


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


NS). The resting anorectal angle was sig-nificantly more obtuse in IFI than controls(130.6 v 114-1, p<0-025) but PAR had noinfluence on resting angles (pre and post:130-6 v 131-8, NS). Although incontinentpatients have obtuse anorectal angle postanal repair is associated with no physio-logical change and results were disappoint-ing.

Action of cisapride on intractable constipa-tion associated with spinal and pelvic nerveinjury

N R BINNIE, P EDMOND, AND A N SMITHI (SpinalInjuries Unit and Castro-intestinal Unit,Western General Hospital, Edinburgh)Interruption of the parasympathetic nervesto the colon, in the spinal canal or in thepelvis, results in lessened colonic motilityand constipation. Simulating nerve stimula-tion by the release of ACH at the myentericplexus should correct the colonic dysmo-tility. Cisapride releases ACH at themyenteric plexus. Oroanal transit time(OATT) has been studied in 15 patientswho related the onset of their constipationto the time of spinal cord injury (n=10)or pelvic surgery for hysterectomy (n=5),before and after oral administration ofcisapride. During the study each patientacted as his/her own control. Beforecisapride, both groups of subjects showeddelayed OATT (spinal group 187±87hours, hysterectomy group 130+76 hours)the passage of the 'Hinton markers' as abolus in 1-7±0-5 stools. With cisapridethere was a significant reduction in OATT(spinal group now 125±73 hours; hyster-ectomy group 81+31 hours) and themarkers passed in 3-2+ 14 stools. Themotility effect is predominantly on thecolon as the small bowel transit was 3-2±1-3hours before and after cisapride 2-8±0-7hours. Faecal water content in stool speci-mens remained unchanged (62.5% and630o).

Stimulation of anterior sacral nerve roots inman: the proximal extent and differentialeffects on colonic motility

N R BINNIE, P EDMOND, G ti CREASEY, AND A N

SMITH (Spinal Injuries Unit and Gastro-intestinal Unit, Western General Hospital,Edinburgh) Brindley introduced sacralanterior root stimulators for bladder controlin paraplegic subjects. Their effects oncolon function have since been studied inspinally injured man. The present study

ascertains the nature and proximal extent ofthe motility response to sacral nerve rootstimulation in six male paraplegic subjects.With endoscopic access, pressure record-ings were taken from the transverse colonand continued distally using conventionalmanometric methods. There was a pressureincrease in response to S2 and S4 (15±10 cmof water) but S3 stimulation caused thegreatest pressure response (45+2() cm ofwater). This response affected all of the leftcolon but in three subjects began 10 cmproximal to the splenic flexure confirmedradiologically. In the lower left colon themotor responses were peristaltic propa-gated distally between three pressure trans-ducers in series. This study establishes thatthe extent of colonic motor response tosacral root stimulation extends from or evenproximal to the splenic flexure and suggeststhat the S3 root is the dominant spinal nerveroot for elicitation of motility responses inthe left colon.

Combined sensorimotor deficit in primaryneuropathic faecal incontinence

J ROGERS, J J MISIEWICZ, AND M M HENRY

(Department of Gastroenterology, CentralMiddlesex Hospital, London) Patients withidiopathic faecal incontinence (IFI) oftenreport that they have no sensation ofimpending or actual incontinence. We havestudied 11 patients with IFI, eight women,mean age (±SD) 56-2±16-9 yrs and ninenormal controls (NC), four women, 56-6±15-9 yrs with techniques of mucosal electro-sensitivity (MES) and rectal distension forthe quantitative assessment of anorectalsensation. Pelvic floor motor function wasassessed by calculating fibre density [(FD),an index of denervation/reinnervation],from single fibre EMG's of the external analsphincter; pudendal nerve terminal motorlatency (PNTML); and anal canalmanometry.

Mucosal electrosensitivity showed a sig-nificant (0.001<p<0-02) sensory deficit inthe upper, 18-4 (5-2-25) mA, median(range) v 5-3 (3-9-6-9); middle, 10-1 (5-7-21-4 v 3-7 (1-7-7-2); and lower, 8-9 (3-1-21.6) v 4-4 (3-1-6-0) thirds of the anal canalin IFI compared with controls. Fibre densityand PNTML were significantly increased inIFI compared with controls [1.73 (1-2-2-18)v 1-38 (1-06-1-69), p<0-05 and 2-5 (1-9)-3 15) ms v 1-95 (1-75-2-15), p<0-0021,indicating a motor-neuropathy. There was aconcommitant decrease in anal canal mano-metry [resting pressure: 40 (40-120) cmH2O v 1()0 (40-120), p<0-05; squeeze

pressure: 60 (0-120) cm HO l' (1(00-21(0),p<0.0)021. Rectal sensation was similar inboth groups. This is the first report of asignificant sensory deficit in the anal canal incombination with a motor neuropathy in agroup of patients with primary neuropathicfaecal incontinence.

Sensorimotor pelvic floor neuropathy: acomparison of diabetes mellitus and idio-pathic faecal incontinence

J ROGERS, J J MISIEWICZ, AND M M HENRY(Department of Gastroenterology atndNutrition, Central Middlesex Hospital,London) Diabetic neuropathy is common,but pelvic floor function has not been invest-igated in this condition. We studied 21diabetics (DM), mean age (range) 55-3(36-72) years, 18 patients with idiopathicfaecal incontinence (IFI), 54.1 (25-83)years and 11 age matched controls. Alldiabetics had peripheral sensory neuro-pathy but none had faecal incontinence.Pelvic floor motor function was assessed bycalculating fibre density (FD) from singlefibre EMG of external anal sphincter;pudendal nerve terminal motor latency(PNTML); and anal canal manometry.Sensation was assessed by balloon disten-sion of the rectum and mucosal electro-sensitivity (MES) of the anal canal. FD wassignificantly (p<0-01) increased in DM andIFI compared with controls [median(range): 1-82 (1-42-2-57) and 2 07 (1-21-2.37) v 1-38 (1-06-1-60))], respectively.Mucosal electrosensitivity showed a signifi-cant (p<0.0l) sensory deficit in the upper,middle and lower thirds of the anal canal inDM and IFI compared to controls, 18-2(3-1-13-5) mA and 11-2 (5-2-25) v 5-3(3-4-6-9); 6-2 (2-7-10-7) and 7() (2-1-21.4)v 3-7 (1-7-7-2); 6-2 (2-9-17.5) and 5-7(8-2-21-6) v 4-0 (2-0-6-2)J, respectively.Pudendal nerve terminal motor latency andmanometry were similar to controls in DM.In IFI PNTML significantly (p<0(01)increased 12-50 (1-80-3-15) ms v 1-95 (1-7-2.25)], and manometry decreased [restingpressure: 40 (20-120) cm water v 90 (40-120), squeeze pressure: 40 (0-160) v 140(50-210)1. Rectal sensation was similar inall groups. This asymptomatic sensorimotorpelvic floor neuropathy is a new finding inDM and is in contrast with the symptomaticneuropathy of IFI.

Rectal prolapse, anterior mucosal prolapse,solitary rectal ulcer. Are they caused byrectal herniation through a weak sphincter?

A1363


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


W M SUN ANI) N W READ (>Sub-Department ofHuimlatn Gastroinitestinal Physiology andNutrition, K Floor, Royal HallamshireHospital, Sheffield) Anorectal pressures atrest, during squeeze and straining to blowup a balloon were measured in femalepatients, six with full thickness rectal pro-

lapse (RP), 12 with anterior mucosal pro-

lapse (AMP), and six with solitary rectalulcer (SRU) and in nine female, age

matched normal subjects. The minimumbasal pressure in RP (9±t)05 cm H.O,mean± SEM; p<0(001), AMP (18±3 cm

H.O; p<0.02), and SRU (20±2 cm H,O;p<0.05) were all significantly lower than innormal subjects (28±6 cm H.O). Thesqueeze pressure was significantly lower

in RP (52+3 cm H2O; p<0(02), AMP(58+7 cm H.O; p<t)0(5), and SRU (52+3cm H.O; p<(0)2) compared with normalsubjects ( 1 15 ±21 cm H.O). When normalsubjects blew up a balloon, anal pressure

(100± 10 cm H,O) increased above rectalpressure (72±7 cm H.O; p<0(05). Incontrast, there werc no significant differ-ences between rectal and highest analpressures in RP (rectal, 95± 17 v anal,93+16 cm H2O), AMP (92±12 cm H,O v

91± 13 cm H.O), and SRU (114+17 cm

H.O 106+ 10 cm H,O). The highest analpressure occurred in the most caudadchannel in normal subjects and in the analchannels nearest the rectum in patients. Theresults suggest anal sphincter weaknessallows the rectum to herniate or prolapsethrough the anal canal and confirms similarfindings in SRU.

Effect of nicardipine on anorectal function innormal controls and IBS patients

W M SUN, A EDWARI)S, AND N W READ (Sub-

Dept of Huiman Gastroinitestinal Physiologyand Nutritioni, K Floor, Royal HallamshireHospital, Sheffield) Nicardipine is a calciumantagonist, which reduces smooth musclecontractility by inhibiting the entry ofcalcium into cells. A randomised doubleblind study was carried out on 18 patientswith irritable bowel syndrome (IBS) and on

1() normal controls to compare the effect ofa single dose (20 mg daily) of nicardipine on

subject's sensations and pressure responseson the anorectum to serial inflation of a

rectal balloon with increasing volume of airand to ingestion of fatty meal. In normalcontrols, nicardipine significantly reducedthe postprandial rate of rectal contractions,though it had no significant effect on theanal and rectal responses to rectal disten-sion with volumes of up to 200 ml. The

threshold volume for perception, desire todefecate and for discomfort or pain were allsignificantly lower in IBS patients than innormal subjects. Nicardipine significantlyincreased these volumes in IBS patients butnot in normal controls, and also reduced thecontractile and symptom response to a mealin IBS patients. The results from this studysuggest that nicardipine could reduce theactivity of colonic smooth muscle in bothnormal controls and IBS patients, and alsoreduce the hypersensitivity of the rectum inIBS.

A histopathological study of severe constipa-tion

N D HEATON, P KLUCK, J R GARRETt, AND E R

[IOWARD (King's College Hospital, Deni-mark Hill, London and Dept Cell Biologyand Genetics, Erasmus University, Rotter-dani, Holland) Resection specimens from20 patients with severe constipation(mean age 14 years; male:female 11:9) notresponding to medical treatment and whohad a partial or subtotal colectomy wereexamined histochemically (simple hydro-lases to identify neurones; nerves assessedfor cholinesterase activity and catechola-mine fluorescence), and immunocyto-chemically for anti-neurofilament antibody.Four distinct histochemical patterns wererecognised - 'normal' (seven patients),aganglionosis (two), hypoganglionosis(six), and hyperganglionosis (five). Neuro-filament 'staining' identified four patternsof immunoreactivity - 'normal' (threepatients), aganglionosis (two), hyper-ganglionosis (one), and 'constipated' (14).There was close correlation between thetwo methods for recognising abnormalcolonic innervation. Four of the sevenpatients, however, with a normal histo-chemical appearance (all slow colonictransit) had abnormal neurofilament 'stain-ing'. The histopathological appearance atthe resection margin related to the clinicaloutcome. The combined use of these twotechniques on resection specimens frompatients with chronic constipation identifiesmore subtle abnormalities of colorectalinnervation than routine histologicalmethods.

Internal anal sphincter (IAS) function inneurogenic faecal incontinence

R J NICHOLLS, D Z lUBOWSKI, D E BURL1EIGH,AND M SWASH (St Mark's Hospital, City

Road, London) There is no information onthe IAS in neuropathic faecal incontinence.Six women with major neurogenic faecalincontinence were studied by anal mano-metry, pudendal nerve terminal motorlatency, external sphincter (EAS) singlefibre EMG, and IAS surface EMG. An IASbiopsy obtained during postanal repair(PAR) was tested in vitro for myogenic tone(response to noradrenaline and isoprena-line) and activation of neural elements(response to electrical field stimulation[EFSI and dimethylphenylpiperazinium[DMPP1). Electron microscopy (EM) wascarried out on each specimen. Sevenpatients undergoing rectal excision wereused as controls.

Pudendal neuropathy was present in allPAR subjects and no controls. Slow waveIAS EMG activity (20-40/min) was presentin 6/7 controls and absent in 4/6 PARpatients. IAS muscle strips from all controlsshowed normal in vitro pharmacologicalresponses whereas in PAR subjects therewas complete insensitivity (four) and slightcontraction to noradrenaline and relaxationto isoprenaline (two). Electron microscopyshowed normal smooth muscle in six con-trols and minor degeneration in one. In allPAR subjects atrophy and necrosis ofsmooth muscle separated by collagen bandswas found. Neurogenic faecal incontinenceis therefore also accompanied by IASdamage.

Experience of lateral puborectalis divisionfor severe constipation

M A KAMM, P R HAWLEY, AND J E LENNARD-JONES (St Mark's Hospital, City Road,London) Patients with severe constipationoften do not 'relax' their pelvic floor duringdefecation. Lateral division of thepuborectalis and upper external sphinctermuscles was performed in an attempt torestore defecatory function in a group ofthese severely constipated patients.The operation was carried out in 18

patients, all women. Mean age was 32.Patients with Hirschprung's disease andsecondary causes of constipation wereexcluded. Fifteen patients had severe idio-pathic constipation with normal bariumenema, and three patients had a mega-rectum or megarectum and sigmoid. Theoperation was performed unilaterally in 12and bilaterally in six; three patients had theoperation performed twice (on the sameside).Three patients reported symptomatic

improvement (one bilateral, two unilateral

A1364


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


division). Improvement did not correlatewith a change in the EMG activity orballoon expulsion.The operation did not alter the mean

stated interval of 11 days between bowelactions. Two patients no longer requiredlaxatives postoperatively. Three patientsshowed a change to puborectalis relaxationon EMG postoperatively and four were ableto expel a balloon only after operation, butthese changes did not correlate withsymptomatic improvement. Mean restinganal canal pressure was unaffected bysurgery, but the maximum squeeze pressurewas significantly decreased from 89-7±12 1(mean±SEM) to 48-5±7-4 (p<0-01). Fourpatients had slight mucous incontinence as aresult of surgery, but none had major soil-ing.The operation produced symptomatic

improvement in only three of 18 patientsbut there was no major morbidity.

Dynamic colonic scanning - defining theregional defect in severe idiopathic constipa-tion

M A KAMM, D G IHOMPSON, J E LENNARD-

JONES, E WAI KER, 1 BINGIIAM, R SOBNACK,

AND N W GARVIE (St Mark's Hospital, CityRoad, London ECI V 2PS and the LondonHospital, Whitechapel, London) We havedeveloped a technique which enables thedelineation of functionally abnormalcolonic regions and applied it to patientswith severe idiopathic constipation.

Six healthy volunteers and seven patientswere studied. A multilumen PVC tube wasswallowed and its progress to the caecummonitored by screening. 99mTc-DTPA wasinjected into the right colon, followed bybisacodyl solution, a powerful colonicstimulant. Transit was monitored by gammacamera, and in four subjects proximalcolonic pressures were simultaneouslyrecorded.

In the controls bisacodyl caused highpressure peristaltic waves associated withrapid transport of isotope to the rectum.Patients showed a spectrum of abnormalresponse, from slow transit involving thewhole colon to slow transit involving the leftcolon only. The range of hepatic flexure torectum transit time was 1-10 mins in thecontrols, and in four patients was 14-25mins while in three patients the isotopedid not reach the rectum by two hours(p<0-01). In addition to delayed movementof the 'isotope head', patients demonstratedimpaired transport of the bulk of the isotopemass.

The normal colon is therefore capable ofgenerating high pressure peristaltic waveswhich correlate with effective transportof colonic contents. There was a clearlydefinable abnormality in all the patients.Assessment of colonic transit in response toa standard chemical stimulus enables thedelineation of functionally abnormalregions.

Discrimination is not impaired by excision ofthe anal transitional zone after restorativeproctocolectomy and ileo-anal anastomosis

M R B KEIGHIIEY AND M C WINSIET (TheGeneral Hospital, Birmingham) The analtransition zone (ATZ) is richly innervatedby sensory nerve endings and is often notinvolved in ulcerative colitis. It has beensuggested that the ATZ should be pre-served in patients having restorativeproctocolectomy and ileoanal pouchanastomosis. We have studied eightpatients before and after ileoanal pouchanastomosis to determine the influence ofexcising the ATZ on anal sensation andability to discriminate gas from solid stoolpostoperatively. Anal sensation wasmeasured in the anal high pressure zone asdefined by manometry using a unipolarconstant current stimulator. The medianthreshold sensation in the lower zone of theanal canal postoperatively was 62 mAcompared with 134 mA postoperatively(p<0-02). Mid zone and upper zone thres-hold values had a median of 67 mA and 77mA preoperatively but sensation wasunrecordable (>150 mA) after restorativeproctocolectomy (p<0.0 1). Despite theobjective loss of sensation in the upper analcanal, only one of the eight patients wasunable to discriminate gas from semisolidstools three months after operation, soilingwas reported in only one patient, nonewere incontinent and all were able todefer defecation for more than one hour.Although the excision of the ATZ impairsanal sensation, functional results do notappear to be impaired.

Aetiology of continued 'defecation' in thepresence of loop ileostomy

M C WINSLE'I', Z DROHL, A ALLAN, J

ALEXANI)ER-WILLIAMS, AND M R B KEIGHLEY

(General Hospital, Birmingham) A loopileostomy is the defunctioning procedure ofchoice but may not guarantee completedefunction. To evaluate this criticism the

defunctioning efficiency of the loopileostomy was assessed in 23 patients. Sixpatients experienced episodes of defecationbut only three had signs of stomal retrac-tion.The defunctioning efficiency was assessed

by a modified radio-isotope and dye tech-nique. Efficiency was assessed in the supineposition to maximise overflow (n=7) orduring unrestricted activity (n= 16).The median defunctioning efficiency

(MDE) in the absence of defecation was99-99°, (99-95-100-0) irrespective of bodyposition. In three patients with defecationbut no stomal retraction, the MDE was99-97% (99-97-1000-). Defecation con-tinued despite split ileostomy formation andall three underwent proctocolectomy foracute distal disease. In three patients withdefecation and stomal retraction, the MDEwas 84-7% (31.16-99-0). After stomarevision the MDE improved to 99-99%(p<0-0)1). In the absence of stomal retrac-tion, the loop ileostomy achieves almostcomplete defunction, regardless of whetherthe proximal limb is dependant, and con-tinued defecation is secondary to activedistal disease. Defecation is only due todistal overflow if there can be associatedstomal retraction and this is cured by stomarevision.

Factors influencing the outcome of restora-tive proctocolectomy

MC WINSLET, R F1.INN, AND M R B KElGilil'Y(The General Hospital, Birminigham)Factors influencing the outcome of restora-tive proctocolectomy in 40) patientsfollowed up for a median of 20 (6-46)months has been assessed by univariate andmultivariate analysis. The variablesassessed included age, sex (F=2 1), previouscolectomy (n= 15), covering ileostomy(n=35), rectal cuff preservation (n= 13),endoanal mucosectomy (n= 15), type ofpouch (S-=, W=8, J=31), diagnosis(UC=30, polyp=3, megarectum =4,Crohn's disease=3), vascular mobilisation(n=20), experience (last 20 cases), andcomplications: pelvic sepsis (n=12), analstenosis (n=9), fistula (n=8), obstruction(n=6). Failure was defined as pouchexcision intubation or continued proximalstoma (n=9). The complication score (t)-8)included: bleeding, sepsis. stenosis, fistulaand obstruction. The functional score(0-12) assessed frequency, use of pads,urgency, soiling, incontinence, discrimina-tion, use of anti-diarrhoeals and dietaryrestriction. Failure occurred in two patients

A 1365


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


with Crohn's disease and seven with ulcera-tive colitis and was significantly morecommon after endo-anal mucosectomy(53% v 4%, p<0).01), rectal cuff preserva-tion (46% v 11l o, p<0)(5), and pelvic sepsis(37% v 4%, p<0.01). The complicationscore was only increased after endoanalmucosectomy (3 v 0, p<0(05). Only pelvicsepsis (10 v 1, p<0.01) and fistulae (10 v 1,p<0.05) influenced the functional score.Hospital stay was significantly prolongedafter preserving the rectal cuff (39 v 23 days,p<0.05), endoanal mucosectomy (40 v 23days, p<0.01), pelvic sepsis (48 v 23 days,p<0.01), and fistulae (49 v 24 days,p<0.05).

Segmental colonic transit in ulcerative colitis- effect of Asian versus Caucasian diet

R C SPILl.ER, H H TAY, D B A SILK, AND J J

MISIEWICZ (Department of Gastro-enterology, Central Middlesex Hospital,Acton Lane, London) Recent reports thatproctitis is associated with delayed colonictransit have not allowed for the knowneffect of dietary fibre. In the present studysegmental colonic transit, three day stoolweights and fibre intake have been assessedin 26 patients, 14 with proctitis, 11 withproctocolitis. Ten were taking an Asian and16 a Caucasian diet. There was no differ-ence between the two diet groups in sexratio or disease distribution and activity asassessed endoscopically. Twenty radio-opaque pellets were ingested at 900 am ondays 1-3. Mean transit in hours was 1 2 x thenumber of pellets seen within each segmenton a radiograph taken at 9 00 am on day 4.

Patients taking a traditional Asian diethad larger stool weights, 278±19 g/24 h,(mean±SEM) n= 10, compared with thosetaking a Western diet, 160±24 g, n=16,p<0-01. Total colonic transit was corres-pondingly faster (12+6 v 32+3 h, p<0)01),correlating inversely with stool weight,r=0-56, p<0O1. The difference in fibreintake (28±4 v 20±2 g) was not significantsuggesting that other factors- for example,spices are responsible for the acceleratedcolonic transit. The extent of disease per sedid not influence transit. The difference intransit between active (12-6±3.2 h, n=10)and inactive disease (31±4 h, n=20) was,however, not significant. Thus in ourpatients diet appears to have a more power-ful influence on colonic transit than theirdisease.

Mechanisms of transport of sodium andchloride and the effects of short chain fattyacids in the human infant colon

H R JENKINS, U SCIINACKENBERG, AND P J

MILLA (Institute of Child Health, London) Amajor function of the colon is the conserva-tion of salt and water which may be aided byshort chain fatty acids. The only previousstudies of the mechanisms involved ininfants have been in vivo where electricalgradients influence ionic movements. Wehave carried out a more detailed study oftransport in isolated human infant colonusing a Ussing Chamber and voltage clampprocedure. Stripped L side colonic mucosa(n=6 pairs) was mounted and bathed inKrebs solution. Under short circuit con-ditions Nat (3-45±1 53 [lmol/h/cm2 mean1 SD) and Cl ((0-63±3.61) were absorbedand a residual ion flux consistent withHCO3 secretion approximates Cl absorp-tion. Short circuit current (3-8±0+28)approximates net Na' movement. Sixtymillimoles acetate increased Na' absorp-tion (3-45+ 153 to 7-74±2-25, p<0(05) by alarge increase in mucosa to serosa flux(7 24±0 92 to 13 55±1 62, p<0.01). Theincreased net absorption of Na+ wasmarkedly reduced by 10 ' M amiloride(7.74±2.25 to 1-75±1 72, p<0.01) whichwas also associated with a marked reductionin tissue conductance. These data clearlyshow that in the infant Na+ is absorbedelectrogenically and C1 electroneutrally inexchange for HCO3 unlike in vivo whereCl moves according to the electricalgradient. Short chain fatty acids favourablyinfluence Na+ salvage and do so via theamiloride sensitive Na' channel. Thusbacterial metabolism of carbohydrate in theinfant colon may be important in the con-servation of salt and water.

Breath hydrogen monitoring in Pneumatosiscoli

M SCHLUP, G 0 BARBEZAT, AND V S CHADWICK(Department of Medicine and WellcomeInstitute, University of Otago, Dunedin,New Zealand) Intestinal gas producingbacteria have been implicated in the patho-genesis of Pneumatosis coli. Previousreports of high hydrogen (H2) contentwithin cysts and breath of some patientsprompted us to investigate two femalepatients (aged 57 and 47 years) with colonicpneumatosis coli presenting with tenesmusand rectal bleeding. In both patients meanfasting breath H2 concentrations wereraised: 75 ppm (range 65-87) and 49 ppm(41-54). A bowel lavage (Golytely, 4 L infour hours) reduced breath H2 from 65 ppmto 15 ppm and from 42 ppm to 1() ppm within12 hours. Subsequent treatment for seven

days with an elemental diet (Ensure, 1500KCal/d) supplemented by Golytely 2 L/d inone patient and activated charcoal 20 g/d inthe other patient resulted in persistently lowbreath H2 with a mean of 13 ppm (3-38)and 12 ppm (5-21) as shown on daily breathH2 monitoring. Symptoms improved andsequential plain abdominal films showedvirtually complete regression of cysts.

This well tolerated treatment resulted in amarked reduction of H2 production prob-ably by depriving H2 producing bacteria ofsubstrate. We assume that this reversesconcentration gradients from cyst to bowelresulting in disappearance of the cysts.

COIORECTAI 11

Colonic cellular proliferation rates innormal subjects: the effect of the faecalstream

M WINSIET, A Al LAN, DENISE YOUNGS, AND

M R B KEIGHLEY (The General Hospital,Birmingham) The role of the faecal streamin maintaining colonic mucosal cellularkinetics was assessed by measuring the cellbirth rate (crypt cell production rate -CCPR) of normal rectal mucosa in 10patients undergoing faecal diversion byileostomy for incontinence (n=6) or protec-tion of an anastomosis (n=4). Rectalbiopsies were obtained before diversion andtwo, six, and 12 weeks afterwards. Theeffect of restoration of intestinal continuitywas assessed at the same time points in ninepatients undergoing closure of ileostomy(post incontinence procedure n=4, postanastomotic protection n=5) CCPR wasassessed by an in vitro stathmokineticmethod with vincristine induced metaphasearrest. Faecal diversion was associated witha significant fall in CCPR from 3-4±1-0prediversion to 1-2±0-4 at two weeks,p<0 05. The CCPR subsequently increasedto 1 7±0+4 at six weeks, and 3-3±1 5 at 12weeks. Restoration of intestinal continuitywas not associated with a significant changein CCPR. Faecal diversion causes a signifi-cant transient reduction in rectal CCPR.These data suggest that the faecal streamhas a tropic effect on mucosal cellularproliferation. The mechanism of thisphenomenon is unclear.

Effect of dietary fibre supplementation oncolonic pH in healthy volunteers

G PYE, J CROMPTON, D F EVANS, A G CLARKE,AND J D HARDCASTILE (Department of

A1366


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Surgery, University Hospital, Nottingham)There is evidence of suggest that intra-luminal pH<6-5 may inhibit the formationof carcinogens and hence be protectiveagainst colorectal cancer. The capacity offibre to lower luminal pH may, therefore,be of significance in preventing colonicmalignancy.The ability of dietary fibre supplementa-

tion to acidify the colon has been studied ineight healthy volunteers. GastrointestinalpH was measured using a pH sensitiveradiotelemetry capsule and portablerecording equipment. Subjects werestudied twice before dietary supplementa-tion, twice during the four week period ofsupplementation of 30 g/day with isphagulahusk (Fybogel) and once afterwards.There was a significant reduction in mean

colonic pH in both the right and left sides ofthe colon (6-5 to 5-8, p<0-001; 7-3 to 6-6,p<0-01) after Fybogel compared with con-trol values. A reduction in the percentage oftime with pH>6-5 was also observed (52%to 21% right; 78% to 54% left). Colonic pHhad returned to near normal within twoweeks of cessation of dietary supplementa-tion.

This study shows that isphagula husk cansignificantly acidify both right and left colonand that these effects are short lived.

Role of in vivo "'technetium labelled redblood cell scintigraphy in lower gastro-intestinal haemorrhage

J F SHARP, J P NEOPTOLFMOS, M NICHOLSON,E M WATKIN, AND D P FOSSARD (Departmentsof Surgery and Radiology, Leicester RoyalInfirmary and Leicester General Hospital,Leicester) Twenty eight patients with majorrectal bleeding had in vivo 'mtechnetiumpertechnetate (395-590 mBQ) labelled redcell scintigraphy in an attempt to identifythe haemorrhage site. Serial scans (0-35hours) were taken using a large field gammacamera (106 counts).Three patients were children. There were

17 male and eight female adults, mean age72-4 years (range 51-91 years). The averagehaemoglobin at presentation was 9-5 g/dl(range 5-4-13-5 g/dl). The mean transfusionwas 8-2 units (range 3-20).Twenty nine scans were performed; 22

were positive and in all these the site ofhaemorrhage was identified either duringthe episode of bleeding (n=14) or subse-quently (n=8). The bleeding site was con-firmed by colonoscopy, arteriography or atlaparotomy. Seven patients had negativescans, no bleeding sites were found but two

patients subsequently had colonoscopicpolypectomy. On average after admissionpositive scans were performed earlier 4-3days (range 0-5-14 days) than negativescans 6-7 days (range 1- 14 days).W'Technetium scintigraphy is a useful

procedure and has a high diagnostic yield. Itallows preoperative identification of bleed-ing sites and should be used early whenbleeding is still occurring.

Left sided colonoscopy as screening pro-cedure for colorectal neoplasia in asympto-matic volunteers >45 years

D P FOI FY, P DUNNE, M O BRIEN, J CROWE, T W

O'CALLAGHAN, AND J R LENNON (LouthHospital, Dundalk, Depts of Gastroenter-ology and Pathology, Mater MisericordiaeHospital, Dublin) Screening colonoscopy tosplenic flexure (left sided colonoscopy) wasoffered to 738 asymptomatic volunteers>45 years in an area with high mortalityrate for colorectal carcinoma. Five hundredwere accepted and examined without pre-medication by an Olympus 100 cm colono-scope after preparation by Fletchers enema.Splenic flexure was reached in 435 patients,descending colon in 49 with averageduration of examination eight minutes.Examination failed in 16 patients. Allpolyps were biopsied and if adenomatous or>5 mm subsequent full colonoscopy/polypectomy was done. Of 180 patients withpolyps, three had carcinoma, 86 adenoma,45 hyperplastic and 35 had miscellaneouslesions. The adenoma group comprised 64men (prevalence 20%) and 22 women (pre-valence 12%). Thirty patients (35%) hadmore than one adenoma and 14 (16%) hadadenomas >1 cm. Full colonoscopy in 94patients showed adenomas proximal tosplenic flexure in nine.Our study shows that of 500 asympto-

matic volunteers >45 years, 0-6% hadmalignant and 17-2% potentially malignantlesions in the left colon. The procedure, leftsided colonoscopy, was acceptable (compli-ance 68%) and well tolerated.

Colonoscopic management of malignantpolyps - an audit

I M CHESNER, S MULLER, M EGAN, J NEWMAN,AND E T SWARBRICK (East BirminghamHospital, Selly Oak Hospital, Birminghamand Newcross Hospital, Wolverhampton)Colonic polypectomy is now acceptedmanagement for polyps, however, its role inthe management of polyps with invasive

malignancy is still subject to debate. Onefactor which may influence the outcome isthe presence of vascular or lymphaticinvasion.We have examined the clinical and histo-

logical features in 7t) malignant polypsremoved at colonoscopy. Thirty six werejudged to have carcinoma in situ, of which35 had no further therapy. One had ananterior resection with no evidence ofmalignancy. There were no polyp relateddeaths during follow up (24-48 months). Ofthe 34 who had invasive carcinoma, 15 werejudged to have had incomplete removaleither endoscopically or histologically, allhad surgery - seven had no evidence oftumour, all alive and well (12-55 months).Eight had tumour present: six Dukes' Aalive and well (24-48 months). One Dukes'B, died 10 months and one Dukes' C alive36 months.

Nineteen were judged to have had com-plete excision and 17 had no further treat-ment. Thirteen alive and well (12-56months) two had surgery. Both had residualtumour with evidence of vascular invasion.These results suggest that invasivecarcinoma can be treated by polypectomyonly if vascular invasion is excluded. If indoubt laparotomy is indicated.

Juvenile polyposis - a precancerous condi-tion

J R JASS, C B WILLIAMS, B C MORSON, AND H1 J RBUSSEY (St Mark's Hospital, City Road,London) Juvenile polyposis coli is inheritedas an autosomal dominant condition andcolorectal polyps number between 5-300.Patients usually present in the seconddecade with profuse rectal bleeding.Although the juvenile polyp is ahamartoma, associated colorectal malig-nancy has been described.One thousand and twenty five polyps

from 85 patients with juvenile polyposishave been examined. Eight hundred andthirty four were typical juvenile polyps; 7t)(8%) showed mild and six (1%) showedmoderate dysplasia. None showed severedysplasia. One hundred and sixty eight weredesignated as atypical juvenile polyps.These were multilobated or papillary andcomprised relatively more epithelium andless stroma than the typical variety. Fifty(30%) showed mild, 25 (15%) moderateand three (2%) severe dysplasia. Theremaining polyps included 21 adenomasand two hyperplastic polyps. Sixteenpatients had a colorectal cancer. Twocancers were confined to the head of a

A1367


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


polyp, one being a typical juvenile polypshowing mild dysplasia and the other avillous adenoma showing severe dysplasia.The mean age of patients with cancer was 34years (range 15-59).

It is concluded that juvenile polyposis is arare but important precancerous conditionand that affected patients require carefulsurveillance and genetic counselling.

Desmosomes and human colorectal cancer

J E MARSTON, D R GARROI), F P PARRISH, ANI)

I TAYILOR (University Surgical Unit, South-ampton General Hospital, Southampton)Reduced adhesion between malignant cellsmay result in some becoming dislodged,thus contributing to metastases. Desmo-somes are adhesive intercellular junctionsbetween epithelial cells. The presence anddistribution of desmosomes in 47 primarycolorectal cancers, eight peritoneal recur-rences, and six liver metastases werestudied using fluorescent staining withantidesmosomal antibody. Desmosomeswere present in all tumours investigated,including metastases. In moderately welldifferentiated tumours the distributionwas identical to uninvolved mucosa. Thefluorescence was confined to opposinglateral membranes, being intense in thesubapical regions and reduced towards thebases of the cells. In poorly differentiatedtumours cell polarity was lost and desmoso-mal staining was evenly distributed aroundthe cells. The desmosomal distribution inprimary tumours and metastases wasessentially similar.The stability of desmosomal junctions to

reduced extracellular calcium was alsostudied. In normal epithelium, junctionsremained intact for at least 1-5 hours:between malignant cells, not only desmo-somes but entire cell contacts were dis-rupted within 30 minutes and resulted inrelease of viable cell clumps. This differen-tial adhesiveness of tumour cells in vitromay be of significance in understanding thedevelopment of colorectal metastases.

Immunohistochemical analysis of cell kineticparameters in colonic adenocarcinomas,adenomas, and normal colons

P JOHNSTON, M O BRIEN, P D)ERVAN, J CROWE,

J l,ENNON, AND D N CARNEY (Departmentsof Gastroenterology, Pathology andOncology, Mater Misericordiae Hospital,Dublin) The K1-67 antigen is a nuclearantigen expressed in the proliferative

phases of the cell cycle (GI, G2, S and M).We used a monoclonal antibody to K1-67and an immunohistochemical assay to studycell kinetics in fresh frozen sections ofnormal colon (12), colonic carcinomas (30)and colonic adenomas (27). The K 1-67score represented the percentage epithelialcell nuclei which stained positively.The K1-67 score ranged f'rom 30 to 95

(mean 58) in carcinomas, 24-78 (mean 47)in adenomas, and one to 20 (mean 7-8) innormal colons. The K1-67 positivity in thenormal colons was confined to the lowerthird to half of the crypt. By contrastadenomas showed positivity throughoutincluding the surface epithelium. Wide-spread staining was also noted incarcinomas.No correlation was noted between K 1-67

score, histologic architecture and degreeof dysplasia in adenomas and similarly nocorrelation was noted between the histo-logic grade or dukes staging in carcinomas.The study of cell kinetics using KI-67

immunostaining in tissue sections providesnew insights to the evolution of colonicadenomas and adenocarcinomas.

Photodynamic therapy (PDT) for colorectalcancer

H BARR, S G, BOWN, ANI) N KRASNER (Th1eGastrointestinal Unit, Walton Hospital,Liverpool and The National Medical LaserCentre, Dept of Surgery, University College,London) Photodynamic therapy is a tech-nique for local tissue destruction with lightafter prior drug sensitisation which hasselectivity for malignant tumours. Animalexperimental studies show that colonictumours necrose and slough whereasnormal colon heals by regeneration withoutrisk of perforation. We treated eightpatients (four men, four women) aged43-88 years with colorectal cancers, inoper-able because of metastases (four), severecardic disease (three), or refusal of surgery(one). Five were below the peritonealreflection (6-13 cm from anus) and threeabove (18-31 cm). Each was photosensi-tised with 2-5 mg/kg haematoporphyrinderivative (HpD) iv and treated two andeight days later with 630 nm red light from adye laser, delivered via a 0-4 mm fibrepassed through a colonoscope and inserted1-5 mm into the tumour. Total energy ateach site was 50 J. The depth of invasion,assessed by endosonography in six patientsranged from 0(7-3-2 cm. The mean depth oftumour removed at each treatment site was0-54±0-28 cm (one week after treatment).

In four patients the total tumour volumewas assessed before and after treatment, thevolume removed was 5-1±2 cc (3-4 treat-ment sites each session). One tumour couldnot be detected endoscopically, endosono-graphically, or on biopsy after treatment(follow up five months). One patient with alarge tumour required two units of bloodtwo days after the second treatment. Photo-dynamic therapy is a promising method oftreatment for small colorectal cancers.

Anal leukoplakia

D R DONALTDSON, J R JASS, AND C V MANN (StMark s Hospital, City Road, London) Analleukoplakia is a rare and ill understoodcondition. The clinical findings and path-ology of 27 patients have been reviewed inorder to assess the results of treatment andthe association of' the condition with analcancer.

Patients presented with pruritus or ananal lump. Ages ranged from 31-79 (mean60 years) and the M:F ratio was 20:7.Examination revealed hard, granular whiteplaques surrounded by moist, thickenedperianal skin. The lesion was either wellcircumscribed or circumferential. Thehistological findings were characterised byhyperkeratosis, acanthosis, 'spiky' down-growth of rete ridges and a band-likechronic inflammatory cell infiltrate at thedermo-epidermal junction. There was asynchronous keratinising squamous cellcarcinoma in nine patients (M:F=5:4, meanage 57 years) which was usually welldiffercntiated. A tenth patient (female)developed a perianal squamous cellcarcinoma I1 years after the diagnosis ofleukoplakia - the patient having defaultedfrom follow up for four years. Patients withand without cancer were treated by surgicalexcision±skin flaps or grafts, but leuko-plakia recurred in the majority of cases,often repeatedly. This study has shown thatanal leukoplakia is frequently associatedwith squamous cell carcinoma and the con-dition recurs despite recourse to aggressivesurgery. Patients need careful follow up andcurrent surgical therapy requires re-evaluation.

Intraoperative pelvic cytology accuratelypredicts those at risk of developing localrecurrence of colorectal cancer

CHRISTINE HAIL, S H SIlVERMAN, JANET

MOORE, H TJHOMPSON, AND M R B KEIGHLEY(The General Hospital, Birmingham) Much

A 1 368


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


attention is currently being paid to theimportance of lateral clearance duringoperation for colorectal cancer in order toprevent local pelvic recurrence. We haveexamined 50 patients using intraoperativecytology to detect residual pelvic diseaseand report the correlation between positivecytology and the incidence of local recur-rence. The first 25 patients were examinedby imprint cytology of tumour bed biopsieswhich were then submitted for histologicalexamination. The second 25 patients werestudied by scrape cytology of the fourquadrants of the pelvis and histology wasconsidered positive if there was tumourpresent in the perirectal fat. Eleven of the50 patients had positive cytology. Two diedin the immediate postoperative period;there were five local recurrences in theremainder at a median follow up of 20months. Of the 39 negative patients, ninewere considered positive on histologicalcriteria, with one local recurrence. Therewere four local recurrences in the 30patients negative on both criteria, a highlysignificant difference (0.5>p>0(25 X2). Weconclude that intraoperative cytology pre-dicts those at risk of developing local pelvicrecurrence. It is more important to detectmalignant disease remaining in the patientthan that removed with the histologicalspecimen.

Endoluminal ultrasound in benign anorectaldisease

A DOBI.F AND A J W SIM (St Mary's Hospital,London) The role of endoluminal ultra-sonography in the management of rectalcarcinoma is well established. This tech-nique has not, however, been evaluated inbenign ano rectal disease. Sixteen patientswith a variety of benign anorectal diseaseshave been studied with the Bruel and Kjaer1846 type ultrasound scanner. The sites ofeight abscesses, three ischiorectal, threeintersphincteric and two extrasphincteric,and a fistula-in-ano were correctly demon-strated. Three solitary rectal ulcers wereshown to be mucosal lesions, as was abenign tubulovillous adenoma. Musculardefects were seen in one patient with acomplete anal sphincter division, in twopatients with colorectal anastamotic defectsand by study of in vitro sphincter division inan abdomino-perineal resection specimen.The results of this pilot study clearly show

the capability of endoluminal ultrasono-graphy in defining the pathological anatomyof benign anorectal conditions. Such defini-

tion will allow accurate diagnosis and per-mit a more rational treatment plan to headopted.

Fistulae in anorectal sepsis: the relevance ofmicrobiology and necessity of repeat exami-nation under anaesthesia

M C WINSILET, N S AMBROSE, AND A ALLAN(Selly Oak, General and Queeni ElizabethHospitals, Birmingham) Isolation of colonicaerobes (CA) or 'gut specific' bacteroides(GSB) from anorectal pus is reported to beindicative of a fistula, and in the absence of ademonstrable fistula is a specific indicationfor repeat examination under anaesthetic(EUA) to reduce recurrence. To confirmthese proposals a retrospective review of103 patients with cultured anorectal pus wasperformed: perianal abscess=65, ischio-rectal abscess=38. Incision and drainage(I&D) was performed in 67 patients, andcombined with EUA in only 36. A fistulawas demonstrated perioperatively in sevenpatients (I&D=4, EUA=3) and in 11 atfollow up (I&D=10, EUA=1), medianthree (1-36) months. CA/GSB were iso-lated in 71 patients (68-90%). Only 12 of 18patients with a fistula had CA/GSB present.Of 59 patients (83%) with CA/GSB but noovert fistula only eight required furthersurgery, with 51 (71-8%) requiring nofurther treatment, median follow up two(2-4) years. Isolation of CA/GSB is sugges-tive but not diagnostic of a fistula. An initialEUA should be performed to exclude afistula but irrespective of the culture asecond EUA is not indicated as mostpatients remain asymptomatic. Occultfistulae are not responsible for the highincidence of recurrent sepsis.

ENDOSCOPY

Diagnostic yield of upper gastrointestinalendoscopy in human immunodeficiencyvirus antibody positive patients

I G BARRISON, S M FOSTER, J R W HARRIS, A J

PINCHING, AND J G WALKFR (Departments ofGastroenterology, Genito- Urinary Medicineand Clinical Immunology, St Mary'sHospital and Medical School, London)There is considerable concern about therisks of endoscopy in human immunode-ficiency virus antibody positive (HiVab)patients and some doubts about the diag-nostic yield. We report the results of a

prospective survey of the indications lorendoscopy, findings, and consequentialchanges in management which hatve occur-red as a result of upper G1 endoscopy in thisgroup of patients.

Approximately 30 HIlVah patients cur-rently aittend the genitourinarv medicineclinic of whom 50 have AIDS. A turther I IlAIDS patients have previously beentreated, and 36 subjects have beeii end(lo-scoped. (19 AIDS). The indications forendoscopy were x-ray negative dyspepsil(16), abnormal barium swallow/meal ( I I),intractable dysphagia ( 10), Lupper gastro-intestinal bleeding (four). Abnormalitieswere found in 14/17 HIlVab cases, (82'%),(x--ray negative dyspepsia 7/9, intractabledysphagia 4/4, abnormal harium meal 3/5).The findings in eight HIVab patients (47(%).resulted in the diagnosis of AID)S; (fourKaposi's sarcoma (KS), two oesophagealcytomegalo virus (CMV) ulceration, onegastric lymphomra, one oesophageal tuber-culosis). The remaining six patients hiadoesophageal candidiasis.

Endoscopic and histologicall confirmiationlof suspected gastrointestinal KS wasobtained in five AIDS patienits, and theremaining 14 subjects had oesophagealciandidiasis (six), bleeding peptic ulcer(three), oesophageal CMV (two). alndnormal examination (three).We conclude that a high diiagnostic yield

is obtained at endoscopy in AID)S andHiVab subjects with upper gastrointestinalsymptoms, leading to changes in mianaige-ment in more than 50," of patientsexamined.

Experience with small bowel endoscopy inKuwait

BASIL Al.-NAKIB, S RAI)IIAKRISIINAN, ILABIBA

IAMIN, S M All, ANI) Y 0MFR (Al-AiniriTeaching Hospital, Ku wait anid Kuwait(ancer Control Centre, Ku wait) Endo-scopic examination of' the small bowelbeyond the ligament of Treitz's is donerarely, because of the relative infrequencyof small bowel disease and lack of a suitableequipment. Over a two year period, wehave done 180 such endoscopic examina-tions in 120 patients. A pediatric colono-scope (Olympus PCF) was used als a jejuno-scope by the push type method, and thejejunum up to 60 cm from the duodeno-jejunal flexure was exatmined with targetbiopsies. Abnormalities were present in 81(69%) patients. In the 19 patients who hadimmunoproliferative small intestinal disease(IPSID), four endoscopic groups could he

A 1369)


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


described. Among the 17 patients who hadsubtotal villous atrophy, the mucosa lookedalmost normal endoscopically in the child-ren, while oedema, diffuse nodularity andulcerations as seen in IPSID was observedin three adults. A large jejunal polyp wasfound to be the source of an obscure gastro-intestinal bleeding. In addition, six patientswho had duodenojejunal Crohn's disease,two with schistosomal jejunitis, and onewith Henoch-Schonlein purpura were diag-nosed by jejunoscopy and histology. Thusdirect examination of the mucosa with biop-sies was of significant value in our region,where a variety of small bowel pathologyexist.

Paediatric upper GI endoscopy - an experi-ence of 531 cases

P K H TAM (INTRODUCED BY PROFESSOR R

SHIELDS) (Division of Paediatric Surgery,Department of Surgery, University of HongKong, Queen Mary Hospital, Hong Kong,present address: Department of ChildHealth, Alder Hey Children's Hospital,Eaton Road, Liverpool) Reports on the useof flexible GI endoscopy in the paediatricpopulation have been scarce. Our experi-ence of 531 examinations on 364 children(age range 1 month-16 years) in the past 12years represents one of the largest seriesavailable and illustrates the advantages ofhaving such a service in a regional centre.The vast majority of our procedures did

not require anaesthesia. Indicationsincluded upper GI bleeding (99), abdomi-nal pain (168), foreign body ingestion (76),vomiting (10), dysphagia (5), reassessment(149), miscellaneous (24). Findings werechronic peptic ulcer (47), gastroduodenitis(59), neoplasm (5), foreign body impaction(19), varices (7), stricture (4), normal (201),recurrent disease (32), healing pathology(114), miscellaneous (59). GI bleeding hadthe highest diagnostic yield (84.8%0).Routine endoscopic assessment of responseto ulcer therapy was valuable in detectingrecurrence (27/47) and guiding treatment.The only pitfall associated with flexibleendoscopy was the failure of detection ofrecurrent tracheo-oesophageal fistulas intwo infants. In this age group, the availableendoscopes are still not optimal for thispurpose. ERCP was successfully performedin 11 children with biliary and pancreaticproblems and gave useful anatomical infor-mation to guide management.Therapeutic procedures (55) included

injection sclerotherapy (32), foreign bodyremoval (19), and endoscopic-guided

bouginage of oesophageal stricture (four).Results were uniformly satisfactory. Inparticular, we have found that stricturemanagement was markedly improved bythe use of the new technique of balloondilatation. We had no morbidity.

Gastric vascular ectasia: nine cases andtreatment by laser photocoagulation

J E COXON AND D F BEC[KILY (PlymotuthG(eneral Hospital, Plymouth, Devon)Gastric vascular ectasia (watermelonstomach) is a cause of chronic severe uppergastrointestinal blood loss. The condition israre with only 17 previously recorded cases.We present nine further cases and suggest:(1) The condition may be more commonthan supposed. In four of our cases thetypical appearances were described ata previous endoscopy but the conditionwas not recognised by the endoscopist.(2) There may be an association withpernicious anaemia. Three of our caseswere receiving B12. The median B12 of theother cases was only 202 (range 163-256).All but one of the patients had a microcyticanaemia at the time of diagnosis but of fivepatients who had a previous unrelatedadmission, four had had an MCV greaterthan 95. One patient had antiparietal cellantibodies. (3) The treatment of choice is adistal gastrectomy, performed successfullyin two of our patients. Four patients refusedor were not fit for surgery and in three ofthese patients the condition was treated bylaser photocoagulation of the affectedportion of the stomach using a NdYAGlaser. Blood loss measured with Cr5' label-led red cells in ml/day was significantlyreduced by photocoagulation (before 140,80, 72-5; after 0)4, 0-56, 1-85).

CO2 v air - an evaluation of discomfort aftercolonoscopy

J A WIlSON, i, J IRVINE, R (GOODACRE, J

O'CONNOR, AND G STE,VENSON (Divisions ofGastroenterology an1d Radiology, McMasterUniversity, Hamiltoni, Ontario, C'aniada)Much of the discomfort after colonoscopy iscaused by air insufflated endoscopically.Carbon dioxide (CO,) is absorbed morerapidly than air from the colon and maycause less pain. Fifty six patients under-going colonoscopy were randomised indouble blind fashion to insufflation with airor CO. Symptoms were evaluated byquestionnaire at one, six, and 24 hours aftercolonoscopy. Abdominal radiographs taken

one hour after colonoscopy were scored forgaseous distension and operator question-naire evaluated adequacy of procedure andtechnical difficulty.

During colonoscopy mean pain score of1-7 for air and 1-8 for CO, were similar. Atsix and 24 hours, gas passed with air wasgreater (p<0()01). At six hours pain scorewas 0 and 0)8 (p<0-001 ) and at 24 hours was0) and ()-7 (p<)-0)( 1) for CO, and air respect-ively. Forty five radiographs were evaluatedin five categories- 16 of 17 with CO, in thefirst three (no gas to slight excess) and 28 of28 (p<0-001) with air in the last two(moderate to severe excess). There was nodifference in operator evaluation for air orCOl.

Insufflation with CO, causes less pain topatients and no extra difficulty to endo-scopists. We recommend its routine use incolonoscopy.

A study of intracolonic hydrogen andmethane concentrations in patients

J CROWE, J LENNON, AND E A (GAiI.AGHFER(Gastrointestinal Uniit, Mater MisericordiacHospital, Dublin, Ireland) Samples of intra-colonic gas were obtained during routinecolonoscopy from three groups of patients-namely, polyethylene glycol (PEG) (n=23,samples=38), phosphate enemia (n=34,samples =41) aind mainnitol (n =4, samples=8). Chromatograph analysis was done oneach sample for nitrogen, oxygen, methane,hydrogen, carbon dioxide and water. Airinsufflation was used during each pro-cedure. In the PEG group, potentiallyexplosive concentrations occurred in six of38 samples. 1-10% relative concentration.In the phosphate enema group, four of 41gas samples had 1-1(0% relative concentra-tion and in the mannitol group two of eightsamples. No sample had potentially explo-sive concentrations of methane. All prep-arations were judged to be good by theendoscopist. All patients had a coexistingoxygen concentration of 50/O or more,necessary for combustion of hydrogen ormethane in their explosive ranges.These results show that pockets of gas

containing potentially explosive concentra-tions of hydrogen may occur in the colonafter preparation with PEG, mannitol orphosphate enema, and that carbon dioxide(an inert gas) insufflation should always beused if diathermy is to be used.

Radial pressures during stricture dilatationwith bougies and balloons

A1370


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


J G C COX, G K BUCKTON, AND J R BENNErr

(Hull Royal Infirmary, Anlaby Road, Hull)It has been suggested that balloon dilatationof oesophageal strictures may be gentlerthan bougienage, and that less force isneeded. To assess this pressure recordingswere done during bougie dilatation in eightpatients with benign oesophageal stricture(mean stricture diameter 7-25 mm). Initialdilatation was carried out with Celestindilators. The dilatation was completed witha size 58 Fr Eder-Puestow bougie speciallymodified with a side hole catheterimplanted at its waist, connected to apressure transducer; pressure recordingswere made throughout the dilatation. Thepressures measured were the final dilatingpressures for each dilatation. The resultsshowed that in five out of eight patients thispressure exceeded 49 kPa (370 mm Hg) andin the three other patients was an average of31 kPa (230 mm Hg).

In 45 strictures (mean diameter 7-6 mm)dilated by a 20) mm balloon, the final dilatingpressure was up to 151 kPa (1133 mm Hg) in39 patients and between 151 and 30)2 kPa(2266 mm Hg) in six patients. Dilatationwas incomplete in one patient in eachgroup.Very high pressures are used in both

forms of stricture dilatation. Balloon dila-tation may be inherently safer and moreeffective as these pressures are applied onlyin a radial direction, but this awaits confir-mation.

Endoscopic sphincterotomy for common bileduct calculi in patients with gall bladder insitu considered unfit for surgery

B R DAVIDSON, J P NEOPTOLEMOS, AND D 1.CARR-ILOCKE (Departments of Surgery andGastroenterology, University of Leicester,Leicester) Endoscopic sphincterotomy (ES)was attempted in 106 patients with commonbile duct (CBD) calculi and gall bladderspresent who were considered unfit forsurgery on the grounds of age and frailtyalone (35%) and/or the presence of majormedical problems (65%). Endoscopicsphincterotomy was successful in 105 cases(99%). Early ES related complicationsoccurred in 21 patients (19-8%). Therewere 12 hospital deaths (11-3%) of whichfive were due to biliary causes (4-7%).Complications were more frequent inpatients in whom initial ES did not clear theCBD (p<002) and mortality was greater inthose without confirmation of CBD clear-ance (p<0()1) unless operated upon.Twelve patients developed gall bladder

complications (11-3%) including five withempyema (4.7%0). Analysis of clinical,haematological and biochemical factorstogether with ERCP findings revealed pre-existing cholangitis to be the only factor ofany value in predicting gall bladdercomplications.These results suggest that an active policy

of CBD stone clearance should be followedand that clearance should be confirmed asfar as possible. Mortality from post-EScomplications might be further reduced byuse of other non-operative interventionaltechniques.

Endoscopic management of large (>2 cm)CBD stone associated with recurrentpyogenic cholangitis

J WC LEUNG, SC SCHUNG, S D MOK, ANI) A K CLi (Combined Endoscopy Unit, Prince ofWales Hospital, The Chinese University ofHong Kong, Hong Kong) The technicaldifficulty and risk involved with endoscopicstone extraction increases with size of thecommon duct stone. Stones >2 cm aretraditionally referred for surgery. Patientswith recurrent pyogenic cholangitis oftenhave large, muddy, calcium bilirubinatestones as the cause of biliary obstruction.From January 1985 to November 1986, wehad 32 patients (M= 12, F=20); age 36-85,mean=69 years), with CBD stones greaterthan 2 cm (greatest diameter 2-0-3-6,mean=2-8 cm). The mean diameter of thedistal CBD was 1-)6 cm; 47% had signifi-cant cardiorespiratory problems. Seventeen(53%) had acute suppurative cholangitisrequiring emergency drainage with a naso-biliary catheter. Stone extraction wasattempted with large (6 cm) baskets andmechanical lithotripsy after sphinctero-tomy. Chemical dissolution was performedwith 1% EDTA solution infused via thenasobiliary catheter.

Sixteen (50(%) had duct clearance after amean of 2-8 sessions (range 1-5) with nomortality. The main reason for failed ductclearance was the lack of space within theCBD to open the basket. Sixteen with failedduct clearance underwent surgery (explora-tion of CBD one, choledochoduodenos-tomy 15) with two deaths.We conclude that a combined use of large

baskets and chemical dissolution with I %EDTA and if necessary, mechanical litho-tripsy can remove CBD stones of >2 cmwith safety.

Endoscopic prosthesis for common bile ductstones

1 M DIAS, S R CAIRNS, P R SALIMON, AND P B

COTTON (Department of gastroenterology,The Middlesex Hospital, London) During athree year period to February 1987, 624patients underwent successful endoscopicsphincterotomy for common bile duct(CBD) stones. Ninety of these patients hadincomplete duct clearance at the firstattempt, principally because of the stonesize, and had one or more endoscopicprosthesis (EP) inserted into the CBD.There were two procedure related deaths,one limited postsphincterotomy haemorr-hage and one failed stent insertion.

Follow up of 31 patients (mean age 71 1years) given EP as a temporary measure wasup to 3-1 months (mean 2-5). All patientssubsequently had complete duct clearance,19 endoscopic and 12 at laparotomy. Therewere no deaths. Twenty five patients (meanage 83-8 years) considered unsuitable forfurther endoscopic or surgical therapywere given EP as definitive treatment andfollowed-up to 37-9 months (mean 18-1).During follow up three patients have diedfrom non-biliary diseases, and only twohave required endoscopic stent changes forcholangitis. There were no biliary relateddeaths.These results suggest that when stones

can not be removed endoscopically: (1)temporary stenting permits further electiveendoscopic or surgical therapy; (2) long-term stent placement may provide a usefultherapy for poor risk surgical candidates.

Open access colonoscopy: a single investiga-tion for suspected colonic neoplasia

J J T TATE AND G T ROYI.F (INTRODUCED BYI TAYLOR) (University Surgical Unit, RoyalSouth Hampshire Hospital, Southampton)It has been suggested that colonoscopyshould be the primary investigation ofoccult or overt colonic bleeding. We reportthe results of providing an open-accesscolonoscopy service to general practitionersfor such patients.

All general practitioners in one healthdistrict were advised of the service by letter.Patients were sent an appointment to attendfor colonoscopy without prior hospitalinterview, sigmoidoscopy or barium enema.

In a 12 month period 138 referrals werereceived from 53 general practitioners.Colonoscopy was performed on 131patients (age 19-83, median 63 years). Sixpatients did not attend and colonoscopy wascontraindicated in one. The entire colonwas examined in 92% and there were nocomplications. Findings were carcinoma in

A1371


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


12 patients (Dukes' stage A =4, B=3,C=5); adenomatous polyps in 22; diverticu-lar disease in 35; colitis in 9 and no abnorm-ality in 53.We conclude that open access colono-

scopy has provided a rapid, safe and com-

plete examination of the large bowel with a

high detection rate for neoplasia (26%) andhas combined treatment (polypectomy)with diagnosis in two thirds of these cases.

Open access fibresigmoidoscopy: a com-

parative audit of efficacy

1. KAIRA, W R PRICE, B J M JONES, AND A N

HlAMLiYN (Department of Gastroenterology,Russells Halland Wordsley Hospital, DudleyHealth District, Dudley) We studied 541open access referrals and compared them to495 hospital initiated fibresigmoidoscopiesover five years. Though the number of openaccess fibresigmoidoscopies doubled over

this period, diagnostic yield remainedunchanged at 40(% and was similar to theyield in hospital initiated procedures.Colorectal carcinoma was seen in 64 openaccess compared to 47 hospital referredpatients, the proportion of Dukes' A lesionbeing similar in both groups (34%S0).Carcinoma was not seen in patients under40 years of age. Fibresigmoidoscopy was

unsatisfactory in 54 (open access 31,hospital 23) and pathology was missed in 12(open access seven, hospital five) patients.There were no significant operator differ-ences.

High diagnostic yields were associatedwith diarrhoea and rectal bleeding (900%),altered blood per rectum (95%) and rectalbleeding associated with pain (100(%). Lowyields were associated with abdominal painas sole cause of referral (0%), constipation(9.3%) or abdominal pain with constipation(8-2%). Diagnostic rate was significantlylow in open access patients under 40 years ofage (19-22%).

Referral was considered justified in 475(87:8%) patients and only 54 (16'6%1o)patients with a normal endoscopy requiredfurther investigations.We conclude that open access fibresig-

moidoscopy is effective and cost efficient.Appropriate patient selection on basis ofage and symptoms can reduce referrals by38% and improve diagnostic yield by 25%without affecting sensitivity.

Fructose and sorbitol malabsorption inpatients with functional bowel disease

J J RUMESSIN AND) [(UDMAND-HOYER (INTRO-

DUCED BY P M CIIRISIIANSiN) (Dept oJ

Sulrgical Gastroenterology, HvidovreUniversity Hospital, Denmark and Depart-ment of Gastroenterologx, and InternalMedicine F, Gentofte Hospital, University of

Copenhagen, Hellerup, Denmark) Theabsorption capacities and the symptom pro-

duction of fructose, sorbitol, fructose-sorbitol mixtures and sucrose was investi-gated by means of hydrogen breath tests in

25 consecutive patients with functionalbowel disease. A 25 g dose of fructose was

malabsorbed by 13 patients, in seven ofthese the absorption capacity was below15 g. In contrast, malabsorption of sucrose

was never seen. Fructose malabsorptioncaused significantly increased abdominaldistress compared with the group of patientswith complete absorption of fructose or

compared with ingestion of sucrose. A 5 g

dose of sorbitol was malabsorbed by eight ofthe fructose malabsorbers. Mixtures of 25 g

fructose and 5 g sorbitol caused significantlyincreased malabsorption and abdominaldistress, and in eight of the fructose mal-absorbers the effect on malabsorption was

more than additive.These results show that malabsorption of

small amounts of fructose, sorbitol andmixtures thereof is significant andsymptom provoking in patients with func-tional bowel disease. The findings may havedirect implications for the dietary guidanceto a major group of patients with functionalbowel disease and raise the possibility ofdefining separate entities in this diseasecomplex.

Current spectrum of intestinal obstruction

Ci P MCEN'I'EE, D PENDIR, r) MULVIN, M

MCCUIIOUGH, S NAIEDEDR, S PARAH, M S

BADURDEEN, V FERRARO, C ClHAM, N GIIAI.11AM,AND P MA'ITHEWS (North Tees General,Darlington Memorial, Bishop A ucklaindGeneral, Hartlepool General Hospitals) Thisprospective study was set up to accuratelydetermine the current spectrum of intestinalobstruction in the UK. During a 12 monthperiod (1 Aug '85-1 July '86) 228 patientsrequired a total of 236 admissions to one

of four neighbouring district generalhospitals serving a total population of591 000- an overall incidence of 39-9 cases

per 100 t)00 per annum. Aetiological factorsincluded adhesions 75 (32%), intra-abdominal malignancy 61 (26%), strangu-

lated hernias 59 (25%), volvulus 10 (4'S),acquired megacolon 6 (4%/O ), pseudo-obstruction 4 (2(%), faecal impaction 6

(30/O%), and miscellaneous 18 (6"%). The peakincidence of obstruction due to adhesions,hernias and intra-abdominal malignancyeach occurred in the eighth decade oflife. Urgent surgery (within 48 hours ofadmission) was performed in 300% ofadhesions (22), 61%o hernias (40), and 30%ointra-abdominal malignancies (18). Overallmortality was 11 '4%o (26 deaths) and post-operative mortality was 12-3% (19 deaths).Twenty two patients required a total of 2993inpatient hospital days as a result ofintestinal obstruction. Strangulated herniasand large bowel malignancy still account forover 50%, of all cases of intestinal obstruc-tion in the UK today. A more aggressive

approach aimed at earlier detection andelective treatment of both conditions shouldbe balanced against the risk of postoperativemorbidity + mortality in what is pre-

dominantly an elderly age group.

Studies on the genesis of colic

M A STCOKES, K J MORIARTY, AND B N

CATCIIPOILt (Departments of Surgery anidMedicine, Universities of Manichester anidWe.stertn A lstralia, Hope Hospital, Salford)'Colic' is a clinical term which has defiedprecise definition. We have examined themotor events in the gut associated withsimulated 'colic'. Two 5-lumen tubesystems, each carrying a balloon, were

used. One lumen inflated the balloon whiletwo sensed pressure proximally and twodistally to it by suitable ports, which thusassessed contractions over 24 or 32 cm ofgut. Sensing channels were water perfused(0(2 ml/min) and recorded via transducerson a rectilinear polygraph. 'Colic' was

induced by balloon inflation followed bysaline bolus injection via the port immedi-ately orad to it. Studies were made of theduodenum, jejunum, ileum and colon,passing the tube system orally, via stomas or

per rectum.(1) Pain was invariably associated with a

rise of baseline pressure, indicating contrac-tion in a gut segment, at least 6 cm andsometimes 32 cm long. (2) With pain, basie-line pressures varied widely betweenpatients. (3) Solitary painless contractionwaves of >25() mm Hg occurred. (4) Dura-tion of pain was similar in both small (44+ 13sec, mean+SD) and large (38±17 sec,mean ±SD) intestine. (5) Location of pain,referable to gut segments, was variable;with increasing severity, pain spread across

FUNCTIONAL BOWEI.

A 1 372


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


the midline. We suggest a definition of'colic' as 'a phasic pain usually lasting lessthan one minute with a slow rise to maxi-mum severity and a similar decline'.

Mouth-to-caecum transit time in man isreduced by a novel serotonin M-receptorantagonist

L MELEAGROS, B KRFYMANN, M A GHATEI, AND

S R BLOOM (Department of Medicine, RoyalPostgraduate Medical School, Hammer-smith Hospital, Du Cane Road, London)The novel selective serotonin M-receptor(5-HT3) antagonist ICS 205-930 (Sandoz)enhances gastric circular muscle contrac-tions in vitro and gastric emptying in vivo inguinea pigs, being 100 and 5() times more

potent than metoclopramide. We thereforeinvestigated its actions in eight fastednormal volunteers (ages 23-33) in a

randomised double blind crossover trial.After an initial rest period ICS 205-930 20mg or placebo were administered by slow ivinjection (time 0). At + 15 min the subjectsconsumed a test breakfast (530 kcal) includ-ing a glass of orange juice containing 25 mllactulose and were studied for a further 18X)min. Mouth-to-caecum transit time was

estimated by breath hydrogen analysis of 5minute 20 ml samples of end tidal exhaledair using an electronic monitor. A sustainedrise in breath hydrogen of 3 ppm or more

was considered significant. Blood samplesfor gut hormone measurements were takenat intervals throughout. ICS 205-930 pro-

duced a significant decrease in mouth-to-caecum transit time (mean±SEM min): ICS205-930 93±8, placebo 121±9 (p<0-05paired t test). Drug administration did notalter the gut hormone profile. We have thusshown that ICS 205-930 significantlyreduces mouth to caecum transit time inman.

Do patients with disordered defecation havea primary personality disorder?

S E FISHER, M R B KEIGHI.EY, K BRECON, V

SMART, AND H ANDREWS (Department ofSurgery, The General Hospital, Birming-ham) It is uncertain whether patients withdisordered defecation have a personalitydisorder. We investigated patients withchronic constipation (n=21) and faecalincontinence (n=29) who were offeredsurgical treatment. Patients were sub-divided into those whose symptoms were

controlled by surgery and those who were

not improved. Patients were investigated

using the HAD scale and the GeneralHealth Questionnaire. Their psychologicalstate was assessed preoperatively and six to12 months postoperatively and the resultscompared with a group of age/sex matchedcontrols (n=5-). Statistical analyses were

performed using the Mann Whitney U twotailed test. Constipated patients had signifi-cantly higher HAD depression scores com-

pared with controls (6 (2-12) v 4 (0-8),p<0)-(5). Patients who were improved(n= 13) by operation could be identifiedpreoperatively by significantly lower HADanxiety scales compared with those whowere not improved (n=8), 8 (3-14) 15(1t)-19), p<0-()2) and by significantly lowerHAD depression scores (4 (2-12) v 7(5-I1), p<0'01). Operation itself, how-ever, had no influence on psychologicalscoring. Incontinent patients did not differfrom the controls using the parameterstested and the group improved by operation(n= 14) could not be identified preopera-

tively.These data indicate that only constipated

patients have a personality disorder andthat those likely to be improved by opera-tion can be identified beforehand.

Diagnosis of allied functional bowel dis-orders using monoclonal antibodies andelectromicroscopy

P PURI, T FUJIMOTO, AND BRIAN LiAKE (C

FEIGHERY) (The Children's Research Centre,Our Lady's Hosp for Sick Children,Crumlin, Dublin 12 and The Hosp for SickChildren, Great Ormond Street, London)There are a number of conditions whichclinically resemble Hirschsprung's diseasedespite the presence of ganglion cells on

rectal biopsy. These conditions have beenclassified on the basis of symptoms undervarious names such as pseudo Hirsch-sprung's disease, chronic adynamic ileusand chronic idiopathic intestinal pseudo-obstruction. There have been no systemicstudies to date to distinguish these condi-tions from one another. We examinedbiopsy and surgical specimens from 19 cases

of allied functional bowel disorder by

enzyme histochemistry (AChE), silverimpregnation and electron microscopy in an

attempt to establish the most appropriatediagnostic procedure. A monoclonal anti-body D7, produced in our own laboratory is

a good marker of the autonomic nervous

system. This unique antibody distinguishesvarious neuronal abnormalities of the bowelexcept abnormalities of the argyrophilplexus which are diagnosed by silver stain-

ing. Six of the 19 cases had a smooth muscle

disorder- for cxample, visceratl myopathy.inclusion bodies etc. which could only bcdiagnosed on electron microscopy. Ourdata suggest that the vast majority of' alliedfunctional bowel disorders can be diagnosedby examining a full thickncss rectal biopsyby immunocytochemistry, silver impregna-

tion and electron microscopy.

GASTROI)UODEINAL

Effect of indomethacin and ranitidine ongastric acidity

R P WAIL, P PRICIIARD, r DANESHMINI), N

SMITH, AND M I ANGMAN (Department oJTherapeutics, University Hospital, Notitig-ham) The non-steroidal anti-inflammatoryagents have been implicated in the patho-genesis and complications of peptic ulccr. Itis unclear which mechanisms are involved,but increased acid secretion could be one.Indomethacin enhances histaminergic andbasal but not food stimulated acid secretion.Prolonged unbuffered acidity is also animportant factor in peptic ulcer treatment.We have therefore investigated the effectsof indomethacin and ranitidinc on noctur-nal gastric acidity in 1i) normal volunteers.Each underwent four randomised studiesafter seven days treatment with indometha-cin (50 mg tds) and/or ranitidine (300 mg at1900 h), or placebo (double-dummy).Acidity was measured on half hourlyaspirates of gastric contents from 1900 h to()800 h. Environmental conditions includingmeals were identical on each study day.Median acidity (range) during the night(220() h-080() h) was 41-7 mmol/l (67-6-25-1) on pla/pla; 39-8 (63-1-24-0) on ind/pla; 0)'36 (21.3-0)-) on pla/ran; 0)77 (43-7-t)-0) on ind/ran. Ranitidine significantlydecreased acidity irrespective of con-comitant indomethacin treatment (p<0)-01,Wilcoxson) and indomethacin alone did notsignificantly affect acidity. This suggeststhat ulcerogenic effects of indomethacin arenot due to increased acidity and that raniti-dine should be effective therapy even inpatients receiving indomethacin.

Isolated duodenal tamponade in the treat-ment of bleeding duodenal ulcer

T V lAYl!OR AND A 1. BLOWER (Department ofSurgical Gastroenterology, ManchesterRoyal Infirmary, Oxford Road, Man-

A1373


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


chester) The overall mortality for bleedingduodenal ulcer remains today, as it hasthroughout this century, at 10%. Endo-scopic methods of arresting haemorrhagehave appeal but none, including laser, hasbeen shown substantially to reduce mor-tality. A simple, low cost, easily availablemethod of isolated duodenal tamponadehas been designed and investigated. Aballoon surrounding a central flexible, butnon-collapsible tube fits over the distal 7 cmof an end-viewing gastroduodenoscope.This can be easily introduced into the duo-denal cap on the initial diagnostic endo-scopy and if active bleeding is present it isinflated by means of a 70 cm tube. Thedevice fits neatly in the cap and cannot bedislocated after removal of the endoscope.A standard 1 cm diameter deep ulcer was

created in the duodenums of four anti-coagulated dogs. A steady rate of blood loss(12 ml/15 minutes) was achieved over fourconsecutive 15 minute periods before crea-tion of duodenal tamponade by the balloon.Haemorrhage was immediately completelyarrested and haemostasis remainedthroughout the rest of the experiment. Twohigh risk patients on ventilators weretreated by tamponade which immediatelyarrested haemorrhage from their duodenalulcers, one had a further small bleed follow-ing removal of tamponade.

Is highly selective vagotomy (HSV) a goodoperation for duodenal ulcer (DU)? A longerlook at the answer

K S DUA, M KORUT}{,. P W BRUNT, AND N A

MATHESON (Departments of Surgery andGastroenterology, Aberdeen RoyalInfirmary, Aberdeen) Highly selectivevagotomy is now well recognised as a treat-ment for DU although the long termefficacy is less well established. We havepreviously resported the early results (at amean of five years postoperatively) of 137patients operated on by one surgeon andrandomly allocated to either HSV or truncalvagotomy and pyloroplasty (TVP). Highlyselective vagotomy was significantlysuperior in Visick grading (p=0-05) andside-effects - for example, diarrhoea,p=-0.01.We have resurveyed the same patients at

a mean of 11 (minimum eight) years post-operatively. There is no difference on blindassessment using rigorous Visick gradingbetween HSV and TVP (Grades 1+11,75-4% and 74 5% respectively). Nearlyone-fifth (18X6%) ofTVP patients (but noneof the HSV patients) had required re-

operation in the intervening period, how-ever- either hemigastrectomy or pyloricreconstruction. The endoscopic provenrecurrence rate for HSV was 5%.Although HSV is a technically more

demanding procedure we regard it as aclearly superior operation to TVP in almostall respects and longterm follow up supportsthis optimism.

Nd-YAG laser therapy in patients withgastric and duodenal malignancies - aninitial experience

J R ANDFRSON AND C G MORRAN (UniversityDepartment of Surgery, Royal Infirmary,Glasgow) Thirty patients (24 men and 6women) of mean age 67 year (range 48-84)with gastroduodenal malignancies werereferred for laser therapy. The reason forreferral was inoperability (n=14) unfit forlaparotomy (n=7), widespread metastases(n=7) and refused surgery (n=2). Thetumour sites were fundus in 12 (involvingcardia in 10), body in four, proximal twothirds in two, distal two-thirds in five,antrum in three, total stomach in three, and2nd part duodenum in one. The majorsymptoms for which relief was sought werebleeding in 21, dysphagia in 11, and vomit-ing from gastric outlet obstruction in eight.On average two laser sessions (range 1-7)were given to each patient.

Relief of symptoms was achieved in 19 ofthe 21 patients with bleeding, 10 of the 11patients with dysphagia with recurrence ofsymptoms needing Celestin intubation infour, and in five of the eight patients withgastric outlet obstruction. Partial relief ofoutlet obstruction was obtained in twopatients who were able to manage fluidsonly. Seventeen of these patients have, todate, died with a mean survival of nineweeks. The mean survival of the remaining13 patients is 28 weeks (range 3-76 weeks).

Palliative laser therapy offers relief ofsymptoms to a significant number ofpatients with gastroduodenal malignanciesunsuitable for surgery, in whom therapeuticoptions are otherwise limited.

Smoking affects mucosal gastrin and gastricacid secretion in duodenal ulcer (DU)

W M HUI, M CHAN, AND S K LAM (Departmentsof Medicine and Clinical Biochemistry,University of Hong Kong, Queen MaryHospital, Hong Kong) Smoking adverselyaffect the healing ofDU and smokers have ahigher maximal acid output (MAO) but the

pathophysiological basis is not known. Toinvestigate this, 30 patients (male. 95% age:35 23+1 79, smokers 33.3%) with activeDU. documented endoscopically wererecruited for the study. The MAO, theparietal cell sensitivity to gastrin (D,"C),fasting and meal-stimulated gastrinresponses were measured. During endo-scopy, three biopsies were taken from thebody of the stomach for the estimation ofhistamine and 2 from the antrum for theestimation of mucosal gastrin. Seven non-ulcer subjects (age 42 7±7-0, male 43%,28-6%YO smoker) were recruited as controls.The results (expressed in mean+SE)showed that MAO was significantly higherin the smokers (32-3±3-5 mmol/h) than thenon-smokers (24 8±1 9) (p<0)05). Themucosal gastrin content was significantlyhigher in the smokers (74.2±16 2 fmol/mg)than non-smokers (41.7±5.9) (p<0(05) andboth were higher than the controls (15.4+5-7) (p<0)01). In both smokers and non-smokers the mucosal histamine (28.3+3.4,28-8+5-1 ng/ml), fasting gastrin level(5.1+0 5, 5 3+0 8 nmol/l), meal stimulatedgastrin response (1.51±0.15 and 1 68±0-29nmol/min/l) and Ds(C (144.6±36.1,156-3±66-3 ng/kg/h) were similar. We con-cluded that the increased MAO in smokersis associated with increase in antral gastrin,suggesting that the latter is important in thepathogenesis.

Basal and stimulated plasma gastrin in duo-denal ulceration (DU): correlation withfailure of H2 receptor antagonists (H2RA)and with recurrent ulceration (RU) afterhighly selective vagotomy (HSV)

K S NAIK, M LOGOPOUIOS, J N PRIMROSE, R l

BLIACKEl71, AND D JOHNSION (UniversityDepartment of Surgery and Anatomy, TheUniversity, Leeds and Leeds GeneralInfirmary, Leeds) The aim was to determinethe influence of gastrin and the proposed'G-cell hyperfunction' in H,RA resistanceand recurrence after HSV in 71 DU patients.Basal (BG) and integrated gastrin responsesto meal (IGRM) and insulin (IGRI) weremeasured before and after HSV for DU.BAO, PAO (Pentagastrin) and PAO(Insulin (I)) were measured pre- and post-operatively. The pre-operative response toH.RA was recorded. Patients were followedup for five years to detect RU. There was nodifference between BG, IGRM and IGRI inH.RA resistant and non-resistant or RU andnon-RU patients. There was a positivecorrelation between preoperative IGRMand postoperative BAO (p= )02). Inte-

A1374


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


grated gastrin responses to meal correlatedinversely with percentage reduction inBAO postoperatively (p=0'02). An inversecorrelation was observed between both pre-operative and postoperative IGRM andPAO(I) (p=0-03). A positive correlationwas shown between postoperative IGRIand postoperative PAO(I) (p=0 03).

G-cell hyperfunction is not an importantcause of resistance of H,RA or of RU afterHSV. Correlations between preoperativeIGRM and postoperative BAO and per-centage reduction in BAO have beendemonstrated. Some of the acid response toinsulin after HSV appears to be mediated bygastrin.

Recurrence of duodenal ulcers duringmaintenance treatment

J PENSTON, D CARTER, AND K G WORMSLEY(Ninewells Hospital, Dundee) The clinicalcourse of 413 patients with endoscopicallyhealed duodenal ulcer has been studied.Thirty seven of the 339 patients (11%)receiving maintenance treatment with anH, receptor antagonist suffered their firstrelapse during the first year; 44/268 (16%)during the second year; 21/168 (13%)during the third year; 11/105 (10%) duringthe fourth year and 3/51 (6%) during thefifth year. Sixty three per cent of firstrelapses were asymptomatic. After endo-scopically confirmed rehealing of the firstrelapse, 50 patients again received main-tenance treatment with 150 mg ranitidine atnight while 46 patients received mainten-ance treatment with the increased dose of300 mg at night. During the first year afterrehealing, 3/46 (7%) patients receiving 150mg ranitidine at night relapsed, comparedwith 0/43 receiving 300 mg ranitidine. Thesubsequent relapse rates were 1/26 (4%)and 1/34 (3%) during the second year offurther maintenance treatment and 2/17(12%) or 1/17 (6%) during the third year,respectively.We conclude that maintenance treatment

is effective in sustaining remission of themajority of patients with duodenal ulcer.The relapse rate during maintenance treat-ment decreased significantly during thecourse of time. Patients who developedrecurrence during maintenance treatmentseemed more likely to remain in remission ifa higher dose of gastric secretory inhibitorwas used for treatment.

Prospective randomised controlled trial ofconservative therapy versus surgery for per-forated duodenal ulcer

T J CROFTS, R J C STEELE, S C S CHUNG, K PARK,AND A K C LI (Department of Surgery, TheChinese University of Hong Kong, HongKong) With the establishment of H.receptor blockers in the treatment for duo-denal ulcer combined with the use of effec-tive antibiotics, we studied if surgical inter-vention is always indicated for patients withperforated duodenal ulcer disease.

Eighty four consecutive patients withdefinite diagnosis of perforated duodenalulcer were randomly allocated to one of twotreatment groups, surgery or non-operative(antibiotics, nasogastric suction and H2receptor blockers) treatment. The types ofsurgery performed depended on the patientfitness and the surgeons' experience. Allconservatively treated patients were

managed by one team and constantlyassessed. If there was no improvement by 12hours after admission, they were referredfor surgery and deemed a failure of con-

servative therapy.There was one trial exclusion leaving 83

patients for analysis; 40 in the conservativegroup and 43 in the surgery group. Bothgroups were comparable for age, sex, ulcer-ogenic factors and presence of coexistantdisease.Ten patients (25%) failed on conserva-

tive therapy but 75% of patients did notneed surgery.

Overall mortality was 4-8% and equal inboth groups. Comparison of morbiditybetween conservative therapy versus

surgery as measured by abscess formation(6 v 2), hospital stay (12 days v 8-8 days) andduration of antibiotic treatment (7.8 days v

5-7 days) all favoured the surgery group.However a subgroup was identified in whichconservative management may be indicatedin preference to surgery.

OESOPHAGUS

Value of omeprazole in the management oferosive oesophagitis refractory to high dosecimetidine

K D BARDHAN, PAMELA MORRIS, MARY

THOMPSON, D S DHANDE, R F C HINCHLIFFE, M

J DALY, N J H CARROLL, AND CATHERINEKRAKOWCZYK (District General Hospital,Rotherham and Astra Clinical ResearchUnit, Edinburgh) Refractory erosive oeso-phagitis (REO) is defined as persistence ofulceration or erosion despite treatment withcimetidine (CIM) 3-2 g daily for :'3 months.To assess if greater acid suppression might

prove more effective in medical manage-ment, 38 patients with REO were treatedwith omeprazole (OME) 40 mg daily for upto eight weeks. Endoscopy, clinical assess-ment and laboratory studies were done atthe start, at four weeks and at eight weeks(if unhealed earlier). Oesophagitis wasgraded 0-4. The demographic featureswere: mean age 61 years; males 50%;smokers 32%; regular NSAID users 11%;obese 47%; mean duration of symptoms 9-9years. Cimetidine treatment, before OMEtherapy, was on average: at 1-6 g daily,18 months; at 2 g, 16 months; at 3-2 g,five months (total 39 months). Endoscopicappearances improved in 28 patients afterfour weeks' treatment and in a further sevenby eight weeks. The cumulative changes ateight weeks were: healed 42%; almost com-pletely healed 45%; unchanged 13%.Symptoms were completely relieved in58%, improved in 32%, unchanged in 80%.Adverse events were infrequent and did notrequire discontinuing OME. There were noclinically significant changes in haema-tology or in biochemistry. Of 28 patientsreturned to maintenance CIM, 50% were inremission after 1-19 months. In conclusion,omeprazole heals or improves oesophagitisand relieves symptoms when refractory tohigh dose cimetidine.

Omeprazole or ranitidine in the treatment ofreflux oesophagitis - result from a doubleblind, randomised, Scandinavian multi-centre study

L LUNDELL, 0 FAUSA, AND S SANDMARK(INTRODUCED BY A WALAN) (Dept of SurgeryII, Sahlgrenska Sjukhuset, Goteborg, Deptof Medicine A, Rigshospitalet, Oslo, andDept of Otorhinolaryngology, Regionsjuk-huset, Orebro, Norway) One hundred andfifty two patients with erosive and/or ulcera-tive oesophagitis were randomised to treat-ment with omeprazole 20 mg, once daily orranitidine 150 mg bid. Endoscopy was doneimmediately before entry and after four andeight weeks' treatment together with assess-ment of symptoms. Macroscopic healing ofoesophagitis was defined as completeepithelialisation of all oesophageal lesions.One hundred and forty four patients com-pleted the first four weeks of treatmentaccording to the protocol and the healingrates were 67% for omeprazole and 31% forranitidine (p=0'00(1). The same greatdifferences in healing rates were seen alsoafter eight weeks treatment; 85°% and 50%for omeprazole and ranitidine, respectively(p<O-000 1).

A 1 375


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Omeprazole gave also a faster and moresubstantial improvement in reflux symp-toms. In the overall evaluation of symp-toms, treatment with omeprazole wassuperior to ranitidine already after oneweek and throughout the study. Heartburnresolved completely in 77% of patientsreceiving omeprazole during the first fourweeks compared to 36% in those receivingranitidine (p<0.-)001). There were noadverse events or clinically significantchanges in the laboratory screen attribut-able to the trial medications. The presentstudy has shown that omeprazole 20 mgonce daily, is superior to ranitidine both inpromoting healing and relieving symptomsin patients with erosive and/or ulcerativeoesophagitis.

Cisapride and gastroesophageal reflux inchildren with or without neurological dis-orders

G S Cl-ARKE, B SANDHU, N GILBERTSON, AND

M J BRUETON (Dept of Child Health, West-minster Children's Hospital, London)Cisapride was used to treat 22 children whohad significant gastroesophageal refluxdemonstrated by 24 hour pH monitoring.Fifteen of these children were neurologic-ally normal (Gp I), seven had major centralnervous system abnormalities (Gp II). Allthe patients underwent repeat pH monitor-ino after three weeks treatment withcisapride. The results were compared usingthe Wilcoxon test. After cisapride treat-ment Gp I showed a significant decrease inmedian % time pH <4 (19 2-9-5;p<001); longest reflux time (36.5-17.2mins; p<0-02); and average reflux duration(7.1-2.1 mins; p<001) indicating a signifi-cant increase in oesophageal clearance rate.

In Gp II the only significant change was adecrease in % time pH >7 (1-8-(-0;p<0.01). In general this group was rela-tively unresponsive to treatment withcisapride. Ranking the responses to treat-ment of both groups showed a high prob-ability that the two groups were respondingdifferently (p>095, Spearman Rankingtest).

Cisapride whose gastrokinetic propertiesare mediated peripherally is effective intreating children with gastroesophagealreflux who do not have central nervoussystem abnormalities.

Comparison of Asilone gel and Gavisconliquid in the treatment of reflux oesophagitis

H 1. SMART AND MICHIAEL ATKINSON (Depart-ment of Surgery, University Hospital,Nottingham) Antacid and alginate ordimethicone combinations are perhapssuperior to simple antacids in treatingoesophagitis. As there is a lack of data onthese combinations we have performed astudy comparing two commonly prescribedpreparations. Fifty three patients withsymptomatic, endoscopically proven oeso-phagitis were randomly allocated to receiveAsilone gel (28) or Gaviscon liquid (25), 10ml qds for eight weeks. Symptom assess-ment and endoscopy were performed at thestart and end of the study. Both groups werecomparable for age and sex distribution,alcohol and tobacco consumption, baselinesymptoms, endoscopy and biopsy scoresand withdrawals (three per group). Aftereight weeks treatment both agents signific-antly (p<0.01) improved symptoms,Asilone gel (79%/ ) being significantly(p<0t)-5) superior to Gaviscon liquid (52%)in relieving heartburn. Only Asilone gel(38%) produced a significant (p<002)improvement in endoscopic oesophagitisfrom baseline scores. Asilone was alsosuperior to Gaviscon (29% v 8% ) inimproving histological scores but this differ-ence was not statistically significant(0-05<p<0.06).The findings of this study suggest that

Asilone gel is superior to Gaviscon liquid inthe treatment of reflux oesophagitis.

Gastric adaptive relaxation in patients withgastro-oesphageal reflux

M N HARTLEY, S J WALKER, AND C R MACKIE

(University Department of Surgery, RoyalLiverpool Hospital, Liverpool) Gastricadaptive relaxation (GAR) may play arole in the pathophysiology of gastro-oesophageal reflux (GOR). We measuredGAR in 13 normal healthy volunteers (HV)and in 12 patients with GOR confirmed by24 h pH monitoring. The HV group were:11 men, two women; mean age 30 years,range 22-41 years; mean body weight 67 kg,range 50 to 75 kg. All were asymptomatic onno medication. The GOR group were: eightmen, four women; mean age 47 years, range23-65 years; mean body weight 78 kg, range60-92 kg. All received no medication 12hours before the study. Endoscopy wasnormal in five, showed hiatus hernia withoesophagitis in four and oesophagitis onlyin three.

Fasted subjects were intubated with aRyle's tube containing a pressure micro-transducer and with a flaccid plastic bag

(800 ml) attached. Gastric corpus funduspressure was recorded during distension ofthe bag with 460±(30 ml (mean+SD) of airover 30 seconds. Pressure indices (median:range) derived from areas under thepressure curves were: HV; 12-7 (7.5-17. 1)cm H,O and GOR; 9-1 (6-4-13.3) cm H,O(p<0(02 Mann Whitney U test). No correla-tion was found between pressure indicesand age, sex or body weight.The result was unexpected. It is, how-

ever, consistent with reports of delayedgastric emptying in patients with GOR.

Patterns of acid reflux in complicated oeso-phagitis

D A F ROBERTSON, M ALDERSLFY, H SHEPHtERD,

AND C L SMI1H (Department of Medicine II,Southampton General Hospital, Southamp-ton) Oesophageal manometry and 24 hambulatory pH recordings from the distaloesophagus were performed in 25 patientswith complications of oesophagitis(stricture, Barrett's oesophagus or oeso-phageal ulcer) and compared with 25patients with uncomplicated oesophagitis.Acid reflux was more severe in the compli-cated group with 26-20% of time below pH 4compared with 11 3% in uncomplicatedpatients (p<0)01). This difference was mostmarked at night, when complicated patientshad long periods of acid reflux with 35 6(Yotime less than pH4 compared with 5-2%.uncomplicated (p<0.001). The mean dura-tion of nocturnal acid reflux was 15-4minutes (2.1 minutes uncomplicated,p<()OO1). Oesophageal motility wasmarkedly abnormal in both groups, but withno demonstrable differences in lower oeso-phageal sphincter pressure (24 mm Hgcomplications, 31 mm Hg oesophagitis, NS)or peristalsis (37% abnormal waves compli-cations, 34-5%/o oesophagitis NS) betweenthe groups. Oesophageal dilatation wasassociated with a non-significant increase inacid reflux (20()3% to 3144%).

Patients with complications of oesopha-gitis have different patterns of acid refluxfrom uncomplicated patients, with pro-longed nocturnal bathing of the oesopha-geal mucosa which may be the cause ofstricture formation, metaplasia or ulcera-tion.

Healing and relapse of reflux oesophagitisafter treatment with omeprazole 20 mg or 40mg daily

D J HETZEL, J DENT, W REED. F NARIELVALA,M MaCKINNON, AND B lIAURENCE (Depts

A 1376


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


of Gastroenterology and Medicine, RoyalAdelaide Hospital, Flinders Medical Centreand Repatriation Hospital, Adelaide, SouthAustralia, and Sir Charles GairdnerHospital, Perth, Western A ustralia)Omeprazole is a highly effective treatmentfor reflux oesophagitis when compared withplacebo. To determine the optimal dose,164 patients with macroscopic erosive orulcerative oesophagitis were randomised ina double blind study to receive omeprazole20 mg or 40 mg daily. Endoscopy wasrepeated at four weeks (and eight weeks ifunhealed) and healing defined as resolutionof all visible erosion or ulceration. Datawere analysed according to intention totreat. At four weeks, healing of oesopha-gitis was significantly better with 40 mg(67/82, 82%) than 20 mg (57/82, 70%,p=-O5), but not at eight weeks (40 mg:70/82, 85%, 20 mg: 65/82, 79%). At day 15symptom relief was slightly faster with 40mg than 20 mg: heartburn was abolished in72% and 57% of patients respectively.

After treatment was stopped patientswere reviewed monthly to determine therate of relapse of ulceration. Endoscopywas repeated on symptom recurrence orafter six months if symptom free. Onehundred and fourteen healed patientsentered this phase. At one, three, and sixmonths, relapse occurred in 39%, 59%, and84% (20 mg) and 39%, 53%, and 89% (40mg). These data show that both omeprazole20 mg and 40 mg daily are highly effectivetherapies for ulcerative oesophagitis butrelapse occurs rapidly when treatment isstopped.

Clinical applications of simultaneous 24hour ambulatory gastric and oesophagealpH monitoring

C S BALL, I L NORRIS, A WATSON, AND T RDeMEESTER (Department of Surgery, RoyalLancaster Infirmary, Lancaster, UK andCreighton University, Omaha, Nebraska,USA) Twenty patients with gastro-oesophageal symptoms underwent com-bined 24 hour pH monitoring of the oeso-phagus and stomach. Twenty asymptomaticvolunteers with normal barium radiologyalso had combined monitoring and pro-vided a control group. Diet, medication,body position and physical activity werecontrolled. Oesophageal acid exposure wascalculated as per cent time spent <pH 4.The effects of meals were excluded from thegastric pH profile and using the remainingdata, gastric alkaline shift (per cent time

>pH 4) and acid shift (per cent time <pH 2)were obtained.

In eight patients the gastric data wereconsidered normal. Six of these hadabnormal oesophageal acid exposure. Theremaining 12 patients had abnormal gastricprofiles. In six, gastric hypersecretion wascorrectly predicted and in one achlorhydriawas diagnosed. The other five patients hadabnormal alkaline shifts suggestive ofduodenogastric reflux confirmed by refluxduring radionuclide cholescintigraphy inthree patients, two of whom had oesopha-gitis but normal acid exposure.Combined pH monitoring is useful in the

early evaluation of upper gastrointestinalsymptoms. It can reliably predict abnor-malities of gastric secretion and can identifypatients whose gastritis and oesophagitismight be associated with reflux of duodenalfluid.

Influence of hiatal hernia on oesophagealfunction

L R JENKINSON, T 1. NORRIS, AND A WATSON

(Department of Surgery, Royal LancasterInfirmary, Lancaster) The imperfectrelationship between gastro-oesophagealreflux (GOR) and the presence of a hiatalhernia (HH) is now well recognised. Con-flicting reports exist, however, on theinfluence of a hiatal hernia of function of theLOS and oesophageal body and thereforeits precise relationship to GOR.

Forty six patients with endoscopic oeso-phagitis underwent a detailed manometricstudy of LOS and oesophageal bodyfunction together with 24 hour ambulatorypH recording. Twenty asymptomaticvolunteers served as controls.Twenty nine patients (63%) had a HH

endoscopically. All patients with oesopha-gitis had a lower median LOS pressure andlonger intra-thoracic length of LOS thancontrols, these changes being independentof the presence of a HH, but patients with aHH had a shorter intra-abdominal LOSlength. The presence of a HH did not affectperistaltic amplitude, but was associatedwith decreased peristaltic velocity in thedistal oesophagus. On pH monitoring, HHwas associated with significantly more refluxepisodes in the supine position andincreased total acid exposure in bothupright and supine positions.We conclude that a hiatal hernia, when

co-existing with reflux oesophagitis,exhibits an effect on the function of both theLOS and the oesophageal body, whichinfluences the pattern of acid exposure.

Cisapride improves oesophageal transit inpatients with dysphagia. Results of arandomised double-blind crossover trial

C A ERIKSEN, D SUTrON, N KENNEDY, AND

A CUSHIERI (Departments of Surgery andNuclear Medicine, Ninewells Hospital andMedical School, Dundee, Scotland) Oeso-phageal dysphagia is a distressing symptomand sometimes difficult to treat. In arandomised double blind crossover trial, weassessed the efficacy of cisapride, a syn-thetic gastrointestinal prokinetic drug, inrelieving dysphagia. Symptomatic patientswith primary motility disorders (n= 13) andreflux oesophagitis (n= 13) were treated fortwo two week periods with cisapride thenplacebo or vice versa. Before and after eachtreatment course, symptom score wasevaluated, and solid bolus oesophageal eggtransit test, 24 hour oesophageal pHmonitoring and endoscopy were per-formed. Although cisapride produced animprovement in symptoms, the effect wasnot statistically different from the placeboeffect. The total oesophageal transit timewas significantly (p<0-008) improved in 17patients (65-4%) after cisapride comparedwith after placebo. There was no significantchange in endoscopic grading after eitherdrug. Oesophageal pH monitoring showedsome improvement in supine acid exposureafter cisapride (pH<4: cisapride 4-2%,placebo 16-5%, mean % data points), butthe differences compared with placebo werenot significant. These preliminary resultsdemonstrate a significant improvement inthe oesophageal transit of dysphagicpatients after treatment with cisapride.

Effects of cisapride (C) upon responses ofthe oesophagus to distension

W G CASE (University Department ofSurgery, The General Infirmary, Leeds)Gastro-oesophageal reflux (GOR) pro-duces oesophageal distension, peristalsisand subsequent acid clearance. Cisapride(Janssen Ltd) reduces the duration ofnocturnal acid reflux in patients with GOR.Does it affect oesophageal responses tointraluminal distension? This has beeninvestigated in 14 healthy male volunteersage 21 years (19-33). With subjects starvedand supine, manometry traces wererecorded from upper oesophageal sphincterand 8 and 14 cm in the oesophageal body bymeans of a Gaeltec catheter passed pernares. The catheter tip in the distal oeso-phagus carried a balloon which was inflatedwith 5 ml water to stimulate oesophageal

A 1 377


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


motility. Frequency, duration and ampli-tude of primary peristalsis (PP), secondaryperistalsis (SP), and tertiary contractions(TC) at 14 cm were quantified for oneminute after balloon inflation under base-line conditions, and again after intravenousinjection of 10 mg C. Traces were analysedblind with regard to treatment. While C hadno significant effects upon PP or TC, SP wassignificantly stimulated, SP/min increasedfrom 0 (0-1) to 1 (0-4)*, with a duration/min of 0 (0-1-8) increasing to 3-3 (0-10-7)sec* and a maximal amplitude increasingfrom 0 (0-20) to 37-5 (0-180) cm HO*after injection of C. The drug may thereforereduce nocturnal acid reflux time by stimu-lation of SP. (All data as median and range,*p<0.01, Wilcoxon signed rank test.)

Manometric findings in patients with theglobus sensation

J C LINSELL, A ANGGIANSAH, AND W J OWEN(Department of Surgery, Guy's Hospital,United Medical and Dental Schools, StThomas St, London) Studies using waterperfused oesophageal manometry systemshave suggested that the globus sensationmay be a manifestation of an oesophagealmotility disorder either of the oesophagealsphincters or of the body of the oesophagus.Using a Gaeltec system of six intra-oesophageal microtransducers we have per-formed manometry on 18 patients (medianage 53 years) whose major symptom was'globus' and who had normal oesophago-scopy and indirect laryngoscopy. Using astation pull through technique the meanlower oesophageal sphincter pressure(LOS) was significantly lower in the globuspatients than in 12 normal controls (p<0-02using Wilcoxon's rank-sum test). Fourpatients had abnormal motility of the oeso-phageal body; two having diffuse oesopha-geal spasm and two non-specific motilitydisorders. There was no significant differ-ence between upper oesophageal sphincterpressures in either globus patients or con-trols. Three patients had abnormalities ofupper oesophageal sphincter function. Ourfindings suggest that patients with theglobus sensation do have oesophagealabnormalities. The finding of hypotensiveLOS pressures suggests a possible associa-tion with gastro-oesophageal reflux.

Dysphagia for solids is a longterm effect ofthe Angelchik antireflux prosthesis

C S ROBERTSON, D L MORRIS, S AMAR, AND J D

HARDCASTLE (Department of Surgery,

University Hospital, Nottingham) Dys-phagia for solids is a common problemfollowing antireflux operations and we havepreviously reported a high incidence afterinsertion of the Angelchik antireflux pros-thesis (ACP). To study this problem wehave performed a marshmallow/bariumswallow (MMBS) in the early postoperativeperiod and at subsequent follow up. Theoesophageal transit time of the marsh-mallow was recorded (normal <1 min). Of40 patients we have initial post operativeMMBS in 36 patients and of these 24 (66%)were prolonged. We have then repeated theMMBS in 21 patients at a median of 33months (9-48) and this long term MMBSwas prolonged in 15/21 patients (71%). Inaddition three patients had the ACPremoved for severe dysphagia before repeatexamination was performed. Of eightpatients with a normal MMBS postopera-tively seven of these remained normal atlong term follow up and of 13 patients withprolonged transit postoperatively, allremained abnormal. We conclude thatoesophageal transit of solids is significantlyslowed by the ACP in many patients andthat this is not a transient postoperativephenomenon.

Expression of normal ranges in 24 hourintra-oesophageal pH monitoring

L R JENKINSON, T L NORRIS, AND A WATSON(Royal Lancaster Infirmary, Ashton Road,Lancaster) The advent of 24 hour intra-oesophageal pH monitoring has revolution-ised the diagnosis of gastro-oesophagealreflux but there is still no uniform methodfor the expression of normal ranges. Datafrom healthy subjects are usually skewedtowards zero and the use of the conven-tional ±2 SD from the mean producesnegative lower limits. We have investigatedthe use of percentiles as an alternativemeans of expression.Twenty six healthy volunteers underwent

manometry and 24 hour ambulatory pHmonitoring. The upper limits of normalitywere estimated from the 90th percentile.Subjects with two or more abnormal valuesin either the upright or supine periods wereclassified as 'refluxers'. Fifty patients withendoscopic oesophagitis were then evalu-ated using this technique. Two (7-7%) con-trols were classified as refluxers (specificity=92%) and four patients as normal (sensi-tivity=92%). The positive and negativepredictive values were 93% and 86%respectively.We conclude that percentiles provide a

useful means of estimating normal rangeswhich demonstrate high sensitivity andspecificity.

Transthoracic colon interposition for benignoesophageal stricture

P E BURKE, E MCGOVERN, AND K M SHAW(INTRODUCED BY G MCENTEF) (Royal City ofDublin Hospital, Baggot Street, Dublin)Thirty four consecutive cases of trans-thoracic oesophageal resection with coloninterposition were reviewed. There were 20men and 14 women ranging in age from twoto 83 years (mean 51.4). Causes of stricturesincluded: reflux oesophagitis (30), corrosiveingestion (two), achalasia (one), andtracheo-oesophageal fistula (one). Themean duration of symptoms was five years,and 18 patients had previous unsuccessfulsurgery. A left thoracotomy with the leftcolon mobilised through the diaphragm wasalways performed. There was one post-operative death (2.9%). Six postoperativecomplications developed (17.6%), onlythree of which were attributable to thethoracotomy. No anastomotic breakdownoccurred. Three recurrent stricturesdeveloped (9%): two oesophagocolic inchildren, which were revised surgically, andone gastrocolic in an adult which required asingle dilatation. Four patients were lost tofollow up, and five were asymptomatic untildeath. Over 70% of patients have beensymptom free during a mean follow upperiod of eight years and all are satisfiedwith the result of their operation.Thoracotomy for oesophageal resectionand left colon interposition is associatedwith few postoperative intrathoraciccomplications and does not predispose toproximal stricture formation in adults. Thehigh level of patient satisfaction confirmsthat the left colon is an excellent visceraloesophageal substitute.

Laparoscopy, ultrasound or CT in the detec-tion of intra-abdominal metastases for oeso-phageal cancer?

I WAll, DOROTHY ANDERSON, I STEWART,G BELL, AND J R ANDERSON (Departmentsof Surgery and Radiology, Glasgow RoyalInfirmary and Inverclyde Hospital,Greenock) A number of non-operativeoptions are now available for the palliationof obstructing carcinomas of the oesopha-gus or cardia. The accurate identification ofintra-abdominal dissemination is thereforeimportant in planning management.

A1378


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


This prospective study evaluates theaccuracy of ultrasound, CT and laparoscopyin 51 patients with tumours of the oesopha-gus or cardia. There were 28 men and 23women, aged 41-93 years (mean 65) ofwhom 26 had squamous and 25 adenocar-cinomas. All patients underwent all threeinvestigations and 43 subsequent laparo-tomy. Eleven patients had histologicallyproven hepatic metastases (seven at laparo-tomy+four laparoscopy/biopsy), 24 hadnodal metastases and five omental/parietalmetastases. Ultrasound correctly identifiedfive hepatic and four nodal metastases (sen-sitivity 45% and 16%); CT six hepatic andseven nodal metastases (sensitivity 54% and29%), and laparoscopy nine hepatic and 13nodal metastases (sensitivity 81% and54%). In addition laparoscopy identifiedfive omental/parietal metastases notdetected by either of the other modalities.Laparoscopy offers a safe accurate means ofassessing intra-abdominal spread of oeso-phageal/cardia tumours and may obviatethe need for surgery in some patients.

Barrett's oesophagus: risk factors formalignancy

P Gil LEN, P KEELING, AND T P J HENNESSY(Department of Surgery, St James'sHospital, Dublin, Ireland) Barrett's oeso-phagus is generally accepted as consequentupon gastroesophageal reflux and has apotential for malignant change. The infre-quency of malignant change makes identifi-cation of causative factors difficult. Toassess whether patients with Barrett's andassociated oesophageal adenocarcinomadiffer from patients with benign Barrett'sepithelium, 42 Barrett's patients - 24 benign(BB) and 18 malignant (MB) have beenstudied over a five year period.The mean ages (65.0 BB v 65-5 MB) and

male:female ratio were similar in bothgroups. Heartburn was significantly morecommon in the benign group (24 BB v 4 MBp<0-001). Dysphagia (10 BB v 16 MB),presence of hiatal hernia (9 BB v 3 MB) andsmoking and alcohol consumption did notdiffer significantly between the two groups.Duration of oesophageal symptoms wassignificantly longer (p<0.001) in the benigngroup (median two years, range 1-20 years)than the malignant group (median eightweeks, range 4-52 weeks).

Dysplastic change was observed inadjacent Barrett's epithelium in six patientswith malignancy and only one in the benigngroup (p<005). Specialised intestinal-typeepithelium was present in all malignant

cases but was seen in only 12 benign cases(p<O 00I).

Malignant Barrett's patients more fre-quently have specialised intestinal typeepithelium and associated dysplastic changewhich, if present in a benign setting, maypredict future malignant change, andshould be closely monitored.

Oesophageal cytology and biopsy: before orafter dilatation?

H J O'CONNOR, A SAVAGE, H TIIOMPSON, R M

VAL ORI, E COLE, AND R COCKEL (Departmentsof Gastroenterology and Histopathology,Selly Oak Hospital, Department of Histo-pathology, General Hospital, Birmingham)Although it is now routine practice to taketissue specimens and dilate oesophagealstrictures at a single sitting, there are nodata on whether cytology and biopsies arebest taken before or after dilatation. In thisstudy, 32 consecutive patients (17 M:15 F,42-89 yrs) with oesophageal stricture(mean lumen size, 4 mm) had brushcytology and biopsies (three) taken beforeand after dilatation (Celestin dilators).Specimens were coded and cytology andhistology assessed blindly by separatepathologists.Of the 32 patients, five had malignant and

27 benign strictures. In the malignantgroup, cytology and biopsy gave one false-negative result each before dilatation butwere diagnostic in all cases after dilatation.There were no false-positive cytology orhistology results before or after dilatationin the benign group. Dilatation had nosignificant effect on biopsy size or on therate of detection of normal, inflamed orulcerated squamous epithelium or necroticdebris. Columnar lined oesophagus was,however, more readily detected after dila-tion, being found in 21 patients comparedwith only 10 before dilatation (x2, p<0.01).Our results suggest that in the evaluation

of oesophageal structures higher diagnosticyield and accuracy is obtained from tissuespecimens taken after dilatation.

Longterm treatment with omeprazole(OME) in resistant refluxoesophagitis:effects on endoscopic healing and fastingserum gastrin (G) levels

E C KLINKENBERG-KNOL, J B M J JANSEN, J W DE

BRUYNE, G F NELIS, H P M FESTEN, P SNEL, A E G

l.UCKERS, AND S G M MEUWISSEN (Depts ofGastroenterology, Free Univ HospitalAmsterdam, State University Leiden,

Slotervaart Hospital, Amsterdam, SophiaHospital Zwolle, Groot Ziekengasthuis DellBosch, RKZ Dordrecht, The Netherlanids)We studied the efficacy and safety of long-term treatment of OME in 35 patients (24M, 11 F, mean age 66 yr. range (R) 42-87)with severe refluxoesophagitis, who wereresistant to treatment for at least 12 weekswith cimetidine 1-6 g or ranitidine 0-6 g. In16 patients complications (bleeding orstrictures) had occurred, eight underwentpreviously antireflux surgery. After 4-12weeks of 4() mg OME od all patients werehealed; then OME was reduced to 20 mg odas maintenance (M). The study includedclinical, endoscopical and biochemicalfollow-ups, including G-levels (every fourweeks in the healing period, every threemonths in M-period). Until May 1987 31/35(89%/) patients were still healed after atleast six months M-treatment (mean 11-7,R:6-16 months). Refluxoesophagitisrecurred in four cases (after 1, 2, 9, and 9months respectively). All four healed againwith OME 40 mg. G-levels (N<100 pg/ml)increased from median 60 (R:6-185) at pre-entry to median 90 (R:10-378) after fourweeks OME 40 mg. In the M-period (OME20 mg) no fall in G-levels occurred: after12 months the median value was 180(R:14-573). No other relevant clinical orbiochemical side-effects occurred.OME-therapy is highly effective. both in

healing as well as in the prevention ofrecurrences of H, receptor resistant reflux-oesophagitis.OME 40 mg produces a significant rise of

fasting G-levels, remaining persistentlyraised up to 12 months despite dose reduc-tion to 20 mg.

Value of endoscopic surveillance in the earlydetection of malignant change in Barrett'soesophagus

C S ROBERTSON, J F MAYBERRY, P D JAMES, ANDM ATKINSON (Department of Surgery andPathology, University Hospital, Notting-ham) Estimates of the risk of adeno-carcinoma developing in Barrett's columnarlined oesophagus vary greatly and most datapreviously presented have been based onprevalence. We present a prospective studyof 56 patients with columnar lined oesopha-gus diagnosed in the 10 year period between1977 and 1986 and followed up with regularendoscopy, and biopsy. Of the 56 patients31 were men, their mean age was 62 years(range 22-84) and the mean follow up wasthree years. During follow up the squamo-columnar mucosal junction was observed to

A1379


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


progress up the oesophagus in 11 patientsbut such progression did not increase therisk of malignancy (x2=0O94; p=ns). Duringthe study eight patients developed severe

dysplasia in the columnar epithelium, twoprogressed to carcinoma in situ and in threepatients an invasive adenocarcinomadeveloped. Diagnosis depended largelyupon biopsy evidence, as in only one of thethree patients was there macroscopic evid-ence of malignancy. These figures indicatethat carcinoma occurred once in every 56patient years follow up, and the relative riskwas 67 times greater than that of the generalpopulation of this age and sex distribution.These findings provide clear evidence thatendoscopic surveillance is important in theearly detection of malignant change compli-cating Barrett's oesophagus.

Palliation of malignant oesophagealstrictures: preliminary comparison ofAtkinson tube and ACMI tumour probe intwo centres

A S MCINTYRE, D L MORRIS, C S ROBERTSON,L SLOAN, J HARRISON, W R BURNHAM, AND

M ATKINSON (Dept Gastroenterology, Old-church Hospital, Essex and Dept Surgery,University Hosp, Nottingham) We havecompared tumour debulking by electro-coagulation with the tumour probe (TP)and Atkinson tube insertion (AT) in 27patients with malignant oesophagealstrictures. Palliation of dysphagia was

assessed subjectively and objectively bybarium swallows+ timed solid swallows (11patients). Thirteen patients received anAtkinson tube (Squam/Adeno/Extrinsic4/6/3) and 14 the tumour probe (5/7/2).Stricturing was upper (AT-1, TP-2), middle(AT-8, TP-2) or lower (AT-4, TP-10).Three patients were transferred from theirinitial treatment groups: one AT to TP(tube displacement), two TP to AT (onerefused repeat endoscopies, one probablyundertreated).

Probe treated patients ate a more varieddiet, though timed solid swallows showedno difference between the groups (95%conf limit AT v TP -36 to 9-0 v -04 to 2-2mm/sec). Six patients required repeatedprobe therapy. Complications seen:

Atkinson tube group- two perforation (onepre-insertion dilatation), two tubes dis-placed, two obstructed by food; Probegroup - one perforation (pretreatmentdilatation), one persistant dysphagia.The tumour probe and Atkinson tube

give comparable palliation and appear com-plimentary. The probe offers an alternative

for high or low oesophageal lesions, thoughrepeated treatment may be necessary.

SMS 201-995 in the management of varicealbleeding

R K MCKEE, S D PRINGLE, 0 J GARDEN, A R

lORIMER, AND D C CARTER (UniversityDepartments of Surgery and MedicalCardiology, Royal Infirmary, Glasgow)Pharmacological arrest of bleeding fromoesophageal varices remains a desirable butelusive goal. A long acting somatostatinanalogue SMS 201-995 has been studied inthis context.

Sixteen stable cirrhotics had wedged andfree hepatic venous pressure measuredusing a Swan-Ganz catheter. Nine patientsgiven a 60 minute intravenous infusion of 25ig/h of SMS 201-995 had a 30% fall intranshepatic venous gradient which was notseen in 7 patients who did not receive thedrug.Twenty two patients with endoscopically

proven active variceal bleeding wererandomised to 48 hours of treatment withoesophageal tamponade or intravenousinfusion of SMS 201-995. Initial control wasachieved in all 11 patients treated withtamponade and nine of 11 patients givenSMS 201-995. Complete control of bleedingover 48 hours was achieved in eight of thetamponade group and six of the SMS201-996 group. These differences are notstatistically significant. SMS 201-995 mayaid temporary control of variceal bleedingwhile more definitive treatment is beingarranged.

Can metoclopramide arrest active bleedingfrom varices? A prospective controlled trial

S W HOSKING, W DOSS, H EL-ZEINY, M S

BARSOUM, AND A G JOHNSON (Kasr El AiniHospital, Cairo and Royal HallamshireHospital, Sheffield) Active bleeding is stillpresent in 30(% of patients reaching hospitalafter variceal haemorrhage. Balloontamponade, vasoactive therapy or immedi-ate sclerotherapy are not always availableor effective and an alternative simplemeasure is needed. Previous studies haveshown that pharmacological constriction ofthe lower oesophageal sphincter (LOS)causes a marked reduction in flow andpressure in submucosal varices proximal toLOS. As 90(% of variceal bleeds occurwithin the zone encircled by the LOS, weinvestigated whether such constrictionwould arrest active variceal bleeding.

Twenty patients with endoscopically provenoesophageal variceal bleeding originatingfrom within 2 cm of the OG Junction, wererandomised to receive either 20 mg meto-clopramide iv or placebo in a double blindmanner. Fifteen minutes after administra-tion each patient was re-endoscoped todocument whether bleeding had stopped.All patients then received sclerotherapy. Ofthe 10 patients who received metoclopra-mide, nine had stopped bleeding comparedto four of the 10 who received placebo(p<O(.01).These results confirm that LOS constric-

tion effectively tamponades bleedingvarices. Metoclopramide may be a simplemethod of arresting variceal bleeding but inpatients requiring transfer to specialistcentres, longer acting drugs may benecessary.

Biphasic diurnal pattern of onset of bleedingfrom oesophageal varices

D SPRENGERS, F D HEYGERE, D COOK, P A

MCCORMICK, N MCINTYRE, AND A K BURROUGHS(Royal Free Hospital School of Medicineand Clinical Epidemiology, London) Overa 42 month period we collected prospect-ively data on 521 separate bleeding episodesin 265 cirrhotics, defined as haematemesisor melaena with a 2 g/dl drop in haemo-globin or change in vital signs. Time of onsetof melaena or haematemesis was recordedaccurately because of our interest in prog-nostic variables for variceal bleeding whichinclude time interval between onset andadmission. The source of bleeding wasoesophageal varices in 417 episodes ofwhich the time of onset was available in 397:114 (29%0) with only melaena and 232(58%) with only haematemesis beforeadmission. The 24 hourly distributionshowed a biphasic pattern with 46% of thebleeds occurring between 7 and 10 am(n= 107) and between 20 and 22 pm (n=76).The Poisson heterogeneity test (Armitage1972: p 214) was used to assess non randomdistribution assuming that the onset ofbleeding would be equally distributedthroughout 24 hours. The asymmetric dis-tribution was significant for the whole group(p<0-001) and for the subgroups: haema-temesis only (p<0-01) and melaena only(p<0001). The time of onset was the samefor patients who started bleeding in hospitalor at home.We do not have an explanation for this

observation but clearly the biphasic diurnalpattern of onset of bleeding from oeso-phageal varices in cirrhotics may have

A1380


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroetterology

relevance to the pathogenetic mechanismsrelated to variceal bleeding.

PANCREA'I'ICO BILIARY POSTERS

Quantitative 99m-Tc DISIDA scanning insphincter of Oddi dysfunction (SOD)

G M FULLIARTON, W R MURRAY, AND T IIILDITCII(Departmenits of Surgery atnd NuclearMedicine, Western Inifirmary, Glasgow)Sphincter of Oddi dysfunction (SOD) is arecognised cause of postcholecystectomypain but is a difficult condition to diagnoscrequiring endoscopic biliary manometry(EBM) to confirm sphincter motor abnor-malities. We have assessed quantitativecholescintigraphy in patients with clinicaland manometric evidence of SOD to deter-mine its value as a non-invasive screeningtcst.

Eight postcholecystectomy (PC) patientswith clinical and manometric evidence ofSOD underwent quantitative 99mTcDISIDA scans (group 1). Eight PC patientswith non-biliary type abdominal pain(group 2) and eight asymptomatic PCpatients acting as controls (group 3) alsounderwent DISIDA scanning. Scinti-graphic analysis demonstrated that the T-max (time in minutes to maximum counts)was significantly increased in group com-pared with both group 2 (25-9±2-1 v150+±1.5, mean+SEM, p<0-(1) and group3 (25-9+2-1 v 130+±0-6, p<0-01). It wasalso noted that the T-max values for group 1did not overlap with either those of group 2or 3, SOD patients always having valuesgreater than the highest group 2 or 3patient.We conclude that quantitative chole-

scintigraphy may be a valuable, non-invasive screening test in clinicallysuspected SOD.

Effects of SMS 201-995 on reticuloendo-thelial system (RES) activity in rats withacute pancreatitis

S A JENKINS, S EILILENBOGEN, [) W DAY,N ROBERIS, ANI) J N BAXTIIR (Departmentsof Surgery, Pathology and Chemical Path-ology, Royal Liverpool Hospital, Liver-pool) The aim of this study was to establishwhether or not some of the beneficial effectsof SMS 201-995 in ameliorating acutepancreatitis are attributable to changes inRES activity. Pancreatitis was induced witha common bile duct ligature in male Wistarrats. Twelve hours later a second ligature

was placed around the bile duct in theregion of the liver hilum. One group of ratsreceived 2 ,tg SMS 201-995 sc and a furtherdose six hours later. A control group weredosed similarly with saline. Reticuloendo-thelial system activity was assessed by thehepatic and splenic uptake ofVTechnetium-sulphur colloid ('Tc-Sc) and circulatinglevels of endotoxin. Pulmonary functionwas assessed by lung compliance, bloodgases and lung weight. The establishment ofpancreatitis was associated an impairmentof RES activity and pulmonary function. Inrats treated with SMS 201-995, RESfunction (liver: blood "Tc-Sc 11-9+1-4,spleen: blood 4-2+0-7) was significantlyhigher (p<()0(5) than in controls (liver:blood 0(4+0-07, spleen: blood 0.2+0.05)and the circulating endotoxin levels signific-antly lower (SMS 201-995 12 3±1-6, con-trols 57-8+4-8 ng/l). Pulmonary functionwas also significantly improved (p<0(05) inSMS treated rats (compliance 0-58±0X06;lung weight 0/o body weight 0.37±0.02)compared with controls (compliance0(83+±0)6, lung weight %, body weight).48±0X03). These results suggest that someof the beneficial effects of SMS 201-995 inameliorating acute pancreatitis may berelated to stimulation of RES activity.

Studies using the CCK-receptor antagonistsCR 1409 point to an important role forcholecystokinin in the stimulation ofpancreatic enzyme secretion by bombesinand a meal

A J 1 DE JONG, N V SINGFR, J B M J JANSE-N, ANI)C B H W LIAMERS (Depts Gastroenterology-Hepatology, Universities of Nijmegen,Essen and Leiden, FRG anid The Nether-latnds) Pancreatic enzyme secretion isassumed to be mediated by both hormonaland nervous mechanisms. The developmentof an in vivo active specific CCK-receptorantagonist (CR 1409) enabled us to deter-mine the role of cholecystokinin (CCK) inthe regulation of bombesin and meal stimu-lated pancreatic enzyme secretion. Six con-scious dogs, equipped with a gastric fistulaand a cannulated pancreatic duct, werestudied for one hour basally and two hoursafter intragastric instillation of a standardtest meal or infusion of 90 pmol/kg/hbombesin or 30 pmol/kg/h cacrulein. Inthese studies either 2 mg/kg/h CR 1409 orsaline were infused in random order.

Peak plasma CCK concentrations weresimilar after the meal and during the infu-sions of bombesin and cacerulein, 7.2±0.8.8X6±0 8 and 8X1±1 2 pM, respectively.

Basal pancreatic protein secretion was notsignificantly affected by CR 1409 (251+64mg/30 min during saline and 163+44 mg/30min during CR 1409). CR 14(19 stronglyinhibited postprandial pancreatic proteinsecretion from 789+ 119 to 44(1+73 mg/30min (p<0.05), bombesin-stimulated pan-creatic protein secretion from 1328+21( to255±37 mg/30 min (p<0(05), and cacrulein-stimulated pancreatic protcin sccretionfrom 1274±236 to 442± 147 mg/30 min(p<0.0 ). The CCK-receptor antagonistdid not significantly influence pancreaticvolume or bicarbonate secretion.These studies show that CCK plays an

important role in the regulation of mcal andbombesin stimulated paincreatic enzymesecretion, but is of little significance for theregulation of basal enzyme secretion and forboth basal and postprandial pancreaticbicarbonate secretion.

Effect of bombesin on serum trypsin aftervarious forms of gastric surgery

J KREUNIN(G, J B M J JANSFIN, AND C B Hi Wl AMERS (Dept Gastroeniterology-Hepatology, University Hospital Leideni,The Netherlands) Infusion of hombesinstimulates several gastrointestinal functionsincluding pancreatic enzyme secretion. Aspostprandial pancreatic enzyme secretion isreduced not only in patients with pancreaticinsufficiency but also in some patients withprevious gastric surgery, we have studiedthe effect of infusion of bombesin (6(0 pmol/kg/20 min) on serum trypsin levels in 1(1patients with truncal vagotomy, nine withparietal cell vagotomy, eight with Billroth1-gastrectomy, 14 with Billroth 11-gastrectomy and 11 with gastrectomy andRoux-en-Y anastomosis, and we have com-pared the results with those of nine patientswith pancreatic insufficiency and 11 normalcontrol subjects. Blood was obtained beforeand at the end of the infusion and serumtrypsin concentrations were measured byradioimmunoassay. Results, expressed asmedian and range, were analysed byWilcoxon's rank sum test.

Incremental serum trypsin in the patientswith parietal cell vagotomy (453; 65-20385[tg/l), Billroth 1-gastrectomy (486; 115-87(1fig/I), Billroth II-gastrectomy (619;51-1002 fig/lI), and gastrectomy with Roux-en-Y anastomosis (317; 19-95(1 fg/I) werenot significantly different from the results inthe normal subjects (697; 199-9975 fig/I).However, in the patients with truncalvagotomy incremental serum trypsin (140;8-43(1 fg/I) was significantly lower

A 1381


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


(p<0(005) than in the normal subjects, butstill higher (p<O(005) than in the patientswith pancreatic insufficiency (0; ()-23 ftg/l).An abnormally low serum trypsin

response to bombesin is found not only inpatients with pancreatic insufficiency butalso in patients with truncal vagotomy. Thelatter finding underlines the importance ofan intact innervation of the pancreas for theserum trypsin response to bombesin.

T'herapeutic fresh frozen plasma (FFP) inacute pancreatitis: intravenous or intra-peritoneal?

A J (GOO[)MAN, N C BIRI), ANI) A G JOHNSON

(University Stirgical Unit, Royval Hallam-shire Hospital, Sheffield) In acute pancreati-tis, plasma and peritoneal levels of X'macroglobulin (x2 M) fall; molecular sizelimits its access to the peritoneal cavitywhere its saturation coincides with death,despite there still being unbound X2 M in thecirculation. Fresh frozen plasma, a sourceOf XYNM was administered either intra-venously or intraperitonieally in a rat modelof acute pancreatitis. Rats were randomlydivided into six groups of 21 animals. Allreceived 14 ml of dextrose/saline by con-tinuous intravenous infusion (IVI) each 24hours. Additionally, groups a, b, c received1t) ml of normal saline, HaemaccelP or ratFFP respectively IVI over 24 hours. Groupsd, e, f received the same 10 ml additions butas a daily intraperitoneal bolus. Theseregimens were repeated for three days.No saline control survived 72 hours.

Intravenous Haemaccel'10 or FFP improvedsurvival (38%, 48%, p<0)-01). Intra-peritoneal Haemaccel® was ineffectivewhereas FFP improved survival further to62°/% compared with saline controls(p<0)-t0)l). Although this was not a signifi-cant improvement over IV FFP, the intra-peritoneal route would allow largervolumes of FFP to be administered.These results suggest two mechanisms for

the improvement: (1) a non-specific effectof iv colloid, (2) supplementation of declin-ing X2 M levels by FFP especially within theperitoneal cavity.

Illness scoring systems - improved precisionfor monitoring the progress of acutepancreatitis?

M lARVIN, AND M J MCMAIION (UniversityDepartment oJ Surgery, The GeneralInfirmlarv, Leeds) Evaluation of potentialtherapy for acute pancreatitis (AP) is

severely limited by the heterogeneousnature of the attacks, wide variation inseverity and the lack of end-points by whichto monitor progress. Systems of multiplecriteria have been developed for prognosis,but are imprecise and provide informationonly during the initial stages of the attack.We have prospectively evaluated three ill-ness scoring systems in 149 attacks of AP,relating the data obtained to clinicalseverity as defined by predeterminedcriteria. Thirty five attacks were classified assevere (11 deaths, 24 major complications),and 114 as mild. Ranson and Imrie scoreswere calculated within 48 hours of admis-sion, and the Acute Physiology-Il Score(APS-II), Simplified Acute PhysiologyScore (SAPS) and MRC Sepsis Score(MRCS) were calculated daily. From dataavailable at 48 hours after admission, theoverall diagnostic accuracy for APS-IT was9(0, Imrie scores 83%, Ranson scores79%, and 78% for SAPS and MRCS scores.At later stages of the attack, dynamicassessment of serial APACHE-IX scorespredicted late complications with a sensi-tivity of 81% and specificity of 96%. Theseresults suggest that the APS-II illness scor-ing system has the potential to provideimproved objective measurement ofseverity and progress of AP.

Experimental studies on interstitial hyper-thermia for treating pancreatic cancer

A C STIEGER, H BARR, R HAWES, S G BOWN, AND

C (iClARK (National Medical Laser Centre,Department of Surgery, University CollegeLondon, The Rayne Institute, London) Mostpancreatic cancers are unresectable at pres-entation and only palliative treatmentis possible. Interstitial low power hyper-thermia has the potential to destroy thesetumours. We studied the nature and evo-lution of local thermal necrosis in normalcanine pancreas produced by the insertionof a 0-4 mm flexible quartz fibre from a lowpower Nd:YAG laser, either up a pancreaticduct or directly into the bulk of the gland.The extent and severity of the necrosisdepended on the laser power and exposuretime used. Some animals developed a lethalpancreatitis, particularly if the duct wasruptured. At I or 1-5 W, lesions up to 1-7 cmin diameter, however, could be producedwhich healed safely.The use of multiple (four) fibres inserted

I cm apart (in a square shape) and activatedsimultaneously a I W each for 1000 secsproduced a lesion up to 7 x 5 cm in size in thehead of the gland, which healed safely with

contraction and scarring of that area of thegland. Hyperamylasaemia was found in allanimals, but, clinical pancreatitis was mild.Pancreatic tumours are more fibrous thannormal gland and so after thermal necrosisare likely to heal better, making this tech-nique worthy of further investigations fortreating pancreatic cancer.

A non-isotopic test for pancreatic insuffici-ency which distinguishes small bowelmalabsorption

I M CHFSNER, J D BERG, AND N LAWSON (EastBirmingham Hospital and SatdwellHospital, West Midlands) We havepreviously demonstrated that para-aminosalicilic acid (PAS) is potentially amore practical substance than 14C-PABAwhen used to derive a PABA/PAS excre-tion index.We have now compared these excretion

indices in 32 subjects including several withsmall bowel malabsorption. In 16 withexocrine insufficiency as judged by the 14C-PABA/PABA ratio the mean (SD) excre-tion index was 14 (8). Using PAS/PABAratio the mean excretion (SD) index was 13(7). There was excellent correlationbetween the two assay methods r=)-92p<O-OI. The sensitivity of the PAS/PABAexcretion index was 95%0 and its specificity85%. Predictive value of test 86%.

In patients with small bowel malabsorp-tion (eight) the mean excretion index wasreduced with a mean (SD) of 55 (13) and thepercentage amounts of PAS and PABArecovered were reduced. Not one of thesubjects with small bowel disease had excre-tion indices as low as those with pancreaticmalabsorption, however.We conclude that the PAS/PABA excre-

tion ratio is a rapid, non-isotopic one daytest of pancreatic exocrine function and candistinguish between pancreatic insufficiencyand small bowel malabsorption. The non-isotopic nature of this test makes it morewidely applicable and repeatable in thesame individual.

Endoscopic ultrasound of pancreatic dis-orders

P J SHORVON, W R .LEES, R A FROST, AND L M

DIAS (Radiology Department, The Middle-sex Hospital, London) Endoscopic ultra-sound (EUS) is performed in a similarmanner to routine upper GI endoscopy, butuses a specially constructed endoscope withan ultrasound transducer incorporated into

A 138X2


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


its tip. Intraluminal bowel gas and bone areby-passed by this technique. By applying ahigh frequency transducer within a fewmillimetres of the pancreas, high resolutionimages of the entire pancreas and ad'jacentorgans are obtained in over 90% of patients.The resolution achieved is better than anyother cross-sectional imaging technique.

Endoscopic ultrasound is employed in thet0-20% of cases of pancreatic disorders inwhich important imaging questions remainafter standard investigations. We have doneover 100 such examinations, and this posterillustrates by case studies the value of thetechnique for each of the six main reasons ofreferral: (1) evaluation of a pancreatic massof uncertain nature, (2) final staging of smallpancreatic neoplasm prior to surgery, (3)localisation of islet cell tumours, (4) evalua-tion of focal pancreatic duct abnormalitiesdiscovered on pancreatography, (5) evalua-tion of early chronic pancreatitis, particu-larly the parenchymal rather than ductalabnormalities, (6) evaluation of low biliaryobstruction of uncertain nature.

In our institution EUS is now the mostspecific imaging test in the diagnosis ofpancreatic disease.

Sex differences in gall stone pancreatitis

T V TAYLOR, S RIMMER, C P ARMSTRONG, S B

l.UCAS, AND J JIEACOCK (Department ofSurgical Gastroenterology, ManchesterRoyal Infirmary, Oxford Road, Man-chester) From a computerised data basecomprising 36 pertinent items in each of 664patients with cholelithiasis, differenceswere studied between men and women. In52 patients there was a documented attackof acute pancreatitis (7-08%). Twenty five of175 men had pancreatitis compared with 27of 489 women (p<0'0001). Although menwith uncomplicated gall stones underwentsurgery at a greater age than women theydeveloped pancreatitis earlier (p<0-00 1). Ahistory of flatulent dyspepsia, chroniccholecystitis and biliary colic was lesscommon in men than women with pan-creatitis (p<0-00l). Men with pancreatitishad fewer stones in their gall bladders thanwomen (X2=9'=8 p=0-002). The commonbile duct and cystic duct in the pancreatiticpatient were more likely to be dilated(p<0-0001); in the non-pancreatitic groupthese ducts were larger in men. Pancreaticduct reflux on operative cholangiographywas more common both in patients withpancreatitis 64% cf 16% (p<0 0001) and inmen (p<0-0001). Men are more susceptibleto gall stone migration at an early stage of

their disease and pancreatic duct reflux.Men have a larger duct system and differentanatomical disposition of the sphincter ofOddi which predispose to pancreatitis.

Computer history analysis in acalculousbiliary pain

G T SUNDERILAND, R P KNILL-JONES, G P CREAN,AND D C CARTER (University Department ofSurgery, Glasgow Royal Infirmary andDiagnostic Methodology Research Unit,Southern General Hospital, Glasgow) Thedecision to investigate acalculous biliarypain (ABP) rests on the history alone,however, the commonly used 'biliaryhistory' has not been defined. We took adetailed history using a computer aideddiagnostic system and compared thefeatures present in patients seen and refer-red for investigation of ABP by generalsurgeons (n=64) with those of confirmedgallstone disease (n=57) and those presentin 'dyspepsia' (n= 1119).

Acalculous biliary pain patients werediagnosed on a 'biliary history' and negativeinvestigations. The decision to operate wasbased on continued symptoms and 25 havecome to cholecystectomy with 85%improved (mean follow up nine months).Of the 39 non-operated patients 381%improved spontaneously, 51% areunchanged and I11% are worse.

Data analysis showed close similaritiesbetween ABP and gallstone patients exceptthat gallstone patients had significantlyshorter histories and suffered more fre-quent 'attacks' of pain. Acalculous biliarypain patients were quite different to'dyspepsia' patients. There were no differ-ences between patients who had chole-cystectomy and those who did not. With theexceptions noted, the symptoms consideredimportant in our population for the diagno-sis of biliary disease are present in thehistory of patients thought to have ABP.

Multivariate analysis of risk factors inpatients undergoing surgery and/or endo-scopic sphincterotomy (ES) for common bileduct (CBD) stones

J P NEOPTOlEMOS, D E SHAW, B R DAVIDSON,AND D L CARR-LOCKE (Departments ofSurgery and Gastroenterology, LeicesterRoyal Infirmary, Leicester) Multivariateanalysis of clinical, laboratory and medicalrisk factors was undertaken in relation topostoperative outcome in 325 patientsundergoing surgery with or without pre-

operative ES for CBD stones (group A) andto post-ES complications in 190 patients(group B). Complications occurred in 83patients (26%) in group A and in 32 (17O)in group B.By univariate analysis age and serum

urea, creatinine, albumin (ALB) andbilirubin (BIL) were all associated withsignificant complications in both groups(p=(003 to <0-001). In group A, however,only BIL, medical risk factors (MRF) andprevious use of ES, were independentlysignificant; the probability of complicationswas logit p-0-51 (MRF)+0.53 (ES)+0(003(BIL) -1-63 - for example, the probabilityof complications with BIL=25 ftmol/l,MRF and ES both=- was p=(- 17 and withBIL=400()mol/l, MRF and ES both= I,this was p=0 637. In group B, only BIL andALB were independently significant; logitp- 1-46- 0-107 (ALB)+0-0()4 (BIL) (forexample with BIL=25 [tmol/l and ALB=40g/l, p=006 and with BIL=400( tmol/l andALB=20 g/l, p=0 68).We conclude that (1) all jaundiced

patients with CBD stones are at risk, (2)patients with medical risk factors should betreated by ES, alone and (3) 'fit' patients arebest treated by surgery.

Ingestion of coffee stimulates plasma chole-cystokinin secretion and gall bladder con-traction in man

B R DOUGt.AS, C W KOCti, J B M J JANSEN, ANI)

C B H W l.AMERS (Depts Gastroenterology-Hepatology and Radiology, UniversityHospital Leiden, The Netherlands) It hasbeen noted that there is an associationbetween coffee drinking and carcinoma ofthe pancreas. Furthermore, ingestion ofcoffee has been held responsible for theinduction of colics in gall stone patients. Ascholecystokinin (CCK) has major stimu-latory actions on both pancreas and gallbladder, we have studied the effect of coffeeon plasma CCK in five healthy volunteers(5 M, age 27-38 yr). After an overnight fastone of three 165 ml isothermic (70°C) andisosmotic (113 mOsm/l) solutions (sodiumchloride, caffeinated and decaffeinatedcoffee) were ingested in random order. Gallbladder volumes were measured by realtime ultrasonography and plasma CCK wasdetermined by a sensitive and specific radio-immunoassay at -5, 0, 10, 20, 30, 40, 50,and 60 minutes.

Ingestion of caffeinatcd coffee inducedan increment in plasma CCK of 2.9± 11pM (p<O-(0) and a rcduction in gall bladdervolume of 11-0±2-0 cm- (p<0'0l). Decaf-

A1383


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


feinated coffee resulted in lower, thoughsignificant (p<0(05), increases in plasmaCCK (1-9±0.8 pM) and reductions of gallbladder volume (5.8± 13 cmI). Ingestion ofthe sodium chloride solution failed to sig-nificantly stimulate plasma CCK release(0(2±0+1 pM) or to induce gall bladdercontraction (3.0±4.0 cm3).

Both caffeinated and decaffeinatedcoffee stimulate secretion of CCK and gallbladder contraction. This effect is not medi-ated by distension or osmotic or thermicfactors. The finding that caffeinated coffeewas more potent than decaffeinated coffeeindicates that not only caffeine but alsoother factors are involved in the coffeeinduced CCK release and gall bladder con-traction.

Sphincter of Oddi motility in acalculousbiliary pain

(i I SUNDERIAND, C G MORRAN, AND D C

CAR I ER (University Department of Surgery,Royal Infirmary, Glasgow) Gall bladdercontraction is considered to be the source ofsymptoms in patients with acalculous biliarypain. Motor abnormalities of the sphincterof Oddi are recognised after cholecystec-tomy but it is not known whether thesedisorders result from surgery or pre-exist.

Using ERCP manometry we haveexamined sphincter motility before chole-cystectomy in 18 patients thought to haveacalculous biliary pain. The papilla of Vaterwas cannulated with a triple lumen catheterand a hydraulic capillary infusion systemused to perfuse each channel with degassedwater. Measurements were made of thefrequency of phasic contractions and ofbasal and peak pressures referenced toduodenal pressure. Eight patients weregiven bolus intravenous injection of chole-cystokinin (I unit/kg) and recordingsrepeated three minutes later. Twelve of the18 patients showed abnormal sphinctermotility. Basal pressure -30 mm Hg wasnoted in five patients. Peak pressures of 300mm Hg and greater were recorded in fourpatients. Five patients had a frequency ofcontractions greater than 8/min and para-doxical responses were observed in five ofthe patients given cholecystokinin. We haveshown for the first time prior to cholecystec-tomy, abnormalities of sphincter motility inpatients with acalculous biliary pain.

Gall bladder contraction in patients withirritable bowel syndrome (IBS)

DAN Z BRAVERMAN (INTRODUCI-D BY M

SININER) (Gastrointestinal Unit, ShaareZedek Medical C(entre, POB 3235,Jerusalem, Israel) Gall bladder contraction(GBC) was measured hy real time ultra-sonography in 20) subjects: eight healthycontrols and 12 with irritable bowel syn-drome (IBS). The technique was used aspreviously described (N Engl J Med 1980;302: 362). The irritable bowel patients wererandomly recruited to the study when it wasevident that they had a complete negativework-up and their complaints were compat-ible with the criteria introduced by Manning(Br Med J 1978; 2: 653). Fasting volumeswere twice as large in the IBS (30(37±3.0ml) as in the control subjects (15.15±0+69ml, p<0-001). Residual volumes were alsotwice as great in those with IBS comparedwith the control subjects (12.91±2-18 ml v5-6+0-58 ml, p<0-(l). Increased fastingand residual gall bladder volumes in the IBSare changes that may promote sequestra-tion of cholesterol or calcium salts in the gallbladder of patients with lithogenic bile.Impaired GBC has been previously shownin pregnant women and in people withdiabetes, coeliac disease, post-vagotomyand cystic duct syndrome. In this studyGBC is shown for the first time to beimpaired in the IBS.

Bile duct diameter after cholecystectomy:assessment by pre and postoperative ERC

S C SCiUNNG, J W C I[UNG, SD MOK, AND)A K CI (Combined Endoscopy Unit, Prince ofWales Hospital, The Chinese University ofHong Kong, Hong Kong) Whether thecommon bile duct increase in size aftercholecystectomy is controversial. We com-pared the greatest diameter of the bile ducton endoscopic retrograde cholangiography(ERC) before and 4-12 months after chole-cystectomy. One hundred and fifty twopatients with uncomplicated gall stonedisease underwent pre-operative ERC.Forty three asymptomatic patients agreedto postoperative ERC 139-398 (mean=243) days after cholecystectomy. None hadcommon duct pathology either on pre-operative ERC or at operation. All hadcholecystectomy alone. Pre and postopera-tive ERC of adequate quality were availablefor scrutiny in 35 patients. The diameter ofthe common duct was measured at its widestpoint. Magnification was corrected for byfactoring down according to diameter of theendoscope measured on the same radio-graph.The greatest diameter of the common

duct measured 0(56-1-58 cm (mean+SD=

0-96±0-27) before and 0-73-1-90 cm(mean+SD= 1-16±0-29) after cholecystec-tomy (p<0-00000l Student's paired t test).The magnitude of dilatation shows a posi-tive correlation with time after surgery(Pearson's correlation coefficient=t)0416,p<0)'02).We conclude that (1) the bile duct on

ERC dilates after cholecystectomy, (2) theprogressive increase in duct size with timesuggests that postoperative dilatation is aphysiological phenomenon rather than overdistension during ERC.

Urgent endoscopic drainage in acutesuppurative cholangitis

J W C LEUNG, S C S CiIUNG, AND A K C LI(Combined Endoscopy Unit, Prince ofWales Hospital, The Chinese Univer.sity ofjHonig Kontg, Hong Kong) Acute suppura-tive cholangitis is caused by bacterial infec-tion of the biliary tract superimposed uponbiliary obstruction, often complicated bysepticaemia and shock. The classical treat-ment is urgent surgical decompression ofthe biliary tract. Between January 1985 andMarch 1987, 100 patients (44 men, 56women; aged 24-95, mean=69 years) withacute cholangitis underwent endoscopicretrograde cholangiography (ERC) duringthe acute illness (mean interval 1-7 days,range six hours to eight days after admis-sion) with a view to urgent endoscopicdrainage. The causes of cholangitis were:common duct stones 85, intrahepatic stoneseight, parasites three, malignant obstruc-tion one, uncertain three. Ninety eight percent of patients had pain, 85% were febrile,800o were jaundiced. Eighty seven per centhad leucocytosis and 96% had derangedLFTs and 40)% were in septicemic shock.The common duct was successfully cannu-lated in 98 patients. The biliary tract wasdrained by sphincterotomy alone (six),insertion of a nasobiliary catheter (26), orboth (65). Endoscopic drainage failed inthree patients who required emergencysurgery and one died. Four died despitesuccessful endoscopic drainage. Endo-scopic duct clearance was achieved in 64patients, while 32 had elective biliarysurgery after the cholangitis settled with onedeath. The overall mortality was 6%.

Urgent endoscopic drainage of the biliarysystem is indicated in patients with acutesuppurative cholangitis.

Role of early ERCP and sphincterotomy inacute biliary pancreatitis

A1384


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of GastroenterologyA

J W C LEiUNG, S C S CIIUNG, S D MOK, ANI) A K C

Li (Cotnbined Endoscopy Unit, Prince ofWales Hospital, The Chinese UniversitvyHong Kong, Hontg Kong) Biliary diseasesare the commonest cause of acutepancreatitis. From November 1984 toNovember 1986, we performed ERCP andsphincterotomy within same hospital admis-sion in 74 patients (M=20. F=54) withacute pancreatitis. the mean age was 56years (range 22-84). Ninety three per centhad hyperamylasemia; 76% had abnormalLFTs. ERCP was performed 7-3 days afteradmission (range 1-25). CBD cannulationwas successful in 70 (95(%O). Biliary diseaseswere identified in 52 patienits (70(3°O): CBDstones (13), IHBD stones (two), ascaris(four), gall stone (33). Three had dilatedCBD only and eight had previous biliarysurgery. No biliary pathology was identifiedin 11 (including four with failed cannulat-tion). The pancreas was filled in 67 patientsand 15 had pancreatitic changes. No exacer-

bation of pancreatitis was observed.Thirteen of the 15 patients with biliarystones (CBD + IHBD stones) had sphincter-otomy and nine had clearance of the CBDstones, two patients had residual intra-hepatic stones. Ascaris were removed endo-scopically in four patients. One patient witha choledochal cyst and three with CBDstones (including two failed sphinctero-tomy) underwent surgery with one death.Thirty of 33 patients with gall stone under-went cholecystectomy during the same

admission with no complication.Early ERCP is safe in acute pancreatitis.

The information obtained allows theoffending biliary pathology to be removedwithin the same hospitalisation.

Ursodeoxycholate treatment and postulatedmechanism of calcification of previouslyradiolucent gall stones

J IIENDERSON (INI'RODUCED BY W PERCY-

ROBB) (Dept of Pathological Biochemistry,Western Infirmary, (lasgow) Medical treat-ment of radiolucent gall stones is not new.

Chenodeoxycholate causes gastrointestinalside effects but not calcification of gallstones; ursodeoxycholate is palatable butcauses gallstone calcification after variableperiods of treatment, precluding furtherdissolution therapy. To investigate a poss-ible mechanism for calcification during dis-solution therapy we have measured carbondioxide solubilisation by ursodeoxycholateand chenodeoxycholate solutions usingtonometry against 5% CO, gas (pCO.35-40 mm Hg). Both bile acid solutions

showed increased CO, solubility: 1 13%zand23% respectively for ursodeoxycholate andchenodeoxycholate (both 150rmM) whencompared with buffer alone.We postulate that during administration

of ursodeoxycholate, the radiolucent stoneis coated with ursodeoxycholate whichavidly solubilises CO, which will be in

equilibrium with HCO3 and CO.,' CO3will form an insoluble precipitate in thepresence of Ca2' ions thus causing calcifica-tion on the surface. Chenodeoxycholatetreatment will coat gall stones with bile acidshowing less avid CO, solubility and is less

likely to act as a source of CO3

Methyl tert-butyl ether for common ductstones: dissolution without radiological dis-appearance

J S DOOILEY, G KAYE, ANi) J A SUMMERFI-iLD

(Royal Free Hospital, London) Methyl tert-butyl ether (MTBE) dissolves cholesterolgall stones rapidly, but common bile ductstones have resisted dissolution (Gastro-enterology 1986; 91: 1296). We have usedMTBE given through a nasobiliary or per-cutaneous transhepatic catheter to treatcommon bile duct stones (10-40 mm dia-meter) in four patients (age 43-87 years).Increasing volumes (0-5-6 ml) were flushedin and out every 0(5 to one hour for sixhours. The session was terminated if thecumulative loss of MTBE reached 10 ml.Patients received one session a day for up tofour days. The maximum volume of MTBEgiven was 80 ml. There was no change instone size on cholangiography in any

patient. Duct clearance was howeverachieved in three patients and in two ofthese, MTBE contributed to the success. Inthe first, the composition of the stones, as

judged by the resistance to basket closure,had changed from rock hard to soft. In thesecond, mechanical lithotripsy, previouslyunsuccessful, crushed one of two largestones. The remaining calculus (30 mm)passed spontaneously.

In summary MTBE solubilises some

common bile duct stones without anychange in their cholangiographic appear-ance. Radiological imaging alone is

insufficient evidence on which to concludethat dissolution has failed.

Enterohepatic circulation patterns in man

D L STOKER AND J G WILLIAMS (Royal NavalHospital, Haslar, Gosport, Hants) No datahave been published on variations in the

enterohepatic circulation (EEHC) in mnan

over 24 hours. This study aims to assess theEHC of bile using a portable giammna probcand SeHCAT. Selected patients (n=8)were given 370 kBq oral SeHCAT. Agamma probe wias secured ovcr McBurney'spoint (terminal ilcum). Counts were

recorded every 15 s for 24 h on a NovoMemolog 2A. Results were analysed by

computer. There were peaks in counts oftwo to four times the mean at between oncand four hours post-SeHCAT ingestion,and these were followed by further peakstwo to four hours later. Further waves thenoccurred with variable frequency, but of a

lower amplitude. In half the patients thecounts dropped below the mean and level-led out during sleep, suggesting that bile

ceased to circulate, and remained in the gyallbladder. In the other half there were con-

tinuing large fluctuations in counts over-

night, despite the fasting state.In conclusion, using an external gamnma

probe and SeHCAT labelled bile, we haveshown that bile conltin1ues to circulate duringsleep in some subjects, while in others itdoes not. This subject merits further study,as major differences in the enterohepaticcirculation may be relavant in a number ofdisease states such as bile malabsorptiondiarrhoea and gall stones.

ILIVER 11

Longterm results of albendazole therapy ofhydatid disease: clinical and experimentalresults

D I MORRIS, M J CiARKSON, K S RICHiARI)S, AND[) IAYILOR (Department of Sulrgery,University Hospital, Nottinghaim) We havepreviously reported encouraging in vitro,animal model and clinical studies of the useof albendazole in hydatid disease. Long-term results, however, are unknown.Naturally infected Welsh sheep underwentradiology to identify cysts, followed bythoracotomy and needle puncture to con-firm viability. Five shcep with live discasereceived albendazole 10 mg/kg by gavaigefor six weeks. Instead of immediatenecropsy, as in earlier studies, we post-poned necropsy for six months to allow anyviable material to recover. At necropsy onlyone of five sheep was found to havc viableprotoscoleces. Electron microscopy showedcomplete destruction of the germinal layerin almost all cysts, but some cysts from thesingle sheep with viable disease had a rela-

A 1385


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


tively normal germinal layer. We have nowtreated 52 patients with albendazole. Of the30 treated in 1982-4 15 initially respondedto chemotherapy, six did not, and nine weretreated prophylactically before or followingsurgery. Of those who responded five havesubsequently recurred and two have beenretreated and have responded again.We conclude that six weeks albendazole

therapy may not be adequate to kill allhydatid cysts but that the majority do notrecur, and those which do may still besensitive to therapy.

Thyroid function as a prognostic indicator incirrhosis

R P BOILTON, L1 SHAPIRO, AND M S IOSOWSKY

(University Department of Medicine andBiochemistry, St James's Hospital, Leeds)Various indices of thyroid function, includ-ing rT3, rT,:T3 ratio and T3:T4 ratio, havebeen proposed as prognostic indicators incirrhosis, but none gives a clear cut indica-tion of progress. Using an easily calculatedindex (Derived Thyroid Index, DTI), basedon a single determination of the rT3:T3 andT3:T4 ratios, we have prospectively evalu-ated 53 patients with cirrhosis of variedaetiology, following patients until death orfor 12 months. Nineteen patients wereChild's grade A, 21 grade B and 13 grade C.Twenty one patients died and a further 14deteriorated during follow up. The DTI was<2-5 in all 18 patients who remained stableor improved, and was >2-5 in 32 of 35(91-40o) of patients who died or deterior-ated. Of 17 patients with a DTI >3-0, 14(82-4%) died, whilst six of 15 patients(40%)) with a DTI between 2-5 and 3-0 died.Derived Thyroid Index was strongly corre-lated with disease severity (r=0-703,p<0'00)1). One patient with a DTI value<2-5 died, and a further two patientsdeteriorated. Thus, a DTI value of >2-5gives a sensitivity of 91 4% and specificity of100% in the prediction of deterioration ordeath during the following 12 months inpatients with cirrhosis, and provides asimple, useful prognostic indicator.

Effect of continued alcohol consumption onbleeding and mortality after varicealhaemorrhage

P A MCCORMICK, I YIN, D SPRENGERS, N

MCINTYRE, AND A K BURROUGHS (Royal FreeHospital School of Medicine, London) Twolarge studies suggest that abstinence aftervariceal bleeding in alcoholic cirrhotics does

not improve survival and one shows thatrebleeding is not affected. We reviewed thesubsequent alcohol history and clinicalcourse of 176 consecutive alcoholic cirr-hotics after their first admission to the RoyalFree Hospital with variceal haemorrhage.Exclusions: died within 30 days 66, absti-nent before bleed six, no follow up four,concomitant other liver disease 6-94patients were followed up: median 20months (range 2-97). Drinking was con-firmed by history, corroboration from rela-tives and blood alcohol levels. Thirtyremained abstinent and 64 continued todrink. Pugh's grading at initial presentationand length of follow up were similar in bothgroups. Mortality was 12/30 (40(%) in theabstinent and 23/64 (36%) in the drinkers.Upper gastrointestinal rebleeding (exclud-ing peptic ulcers) occurred in 53OO (16/31) ofthe abstinent v, 75%o of the drinkers (48/64).There was no significant difference in therebleeding or mortality rates between thetwo groups (logrank test), however, therewas a trend towards a reduction in re-bleeding after eight months of abstinence.Although abstinence probably has a bene-ficial effect this does not occur early enoughto alter the prognosis after variceal bleedingin alcoholic cirrhotics.

Oral urea in the treatment of hyponatraemicascites

P C HAYFS AND ROGER WILLIAMS (Liver Unit,King's College Hospital, London) Theeffect of oral urea in patients with cirrhosisand hyponatraemic ascites was studied in 10patients. Conventional diuretic therapy wasdiscontinued once serum sodium fell below130 mmol/l and patients randomised toreceive three days of oral urea (20 G tid)'early', after the baseline 24 hours, and'late', after a further three days withouttreatment. Weight fell by a mean of 18±2 1kg (SD) in all patients on urea and thisreduction was significantly greater in thoserandomised to urea early than those receiv-ing no therapy over the first three dayperiod (p=0-0)4). Serum sodium rose by4-9±5-3 mmol/l from baseline on ureaalthough the rise was not significantlygreater compared with no therapy overthree days. Urine output measured over thethree day period before and during ureatherapy increased from 718±202 ml to1311+315 ml. Serum urea rose by 14-3±9-6mmol/l mean from baseline and was signifi-cantly higher compared with no therapy(p=0 )03). Serum creatinine fell by 8-2+6-9[tmol/I and urine osmolality increased by

139±145 omol/l compared with beforeurea. The serum urea fell rapidly after ureawas discontinued.

Diarrhoea and vomiting commonlyaccompanied urea therapy, requiring areduction in dosage. Encephalopathyoccurred in one patient. Oral urea acts as anosmotic diuretic producing fall in weightand improvement in hyponatraemia butwithout clinical benefit.

Surgical management of bleeding oeso-pageal varices: Warren shunt versuslienorenal shunt - 15 years' experience

J P A lODGE, A I D MAVOR, AND G R GILES

(University Department of Surgery, StJames's University Hospital, Leeds) Retro-spective analysis of patients undergoinglienorenal shunt (28) or distal splenorenal(Warren) shunt (53) surgery (1971-1986)revealed important predictive factors forsurvival, but showed no significant survivaldifferences between the two shunt groupsover a period of up to 14 years.

In 52 patients (64%) active haemorrhagewas the reason for operation and 17 of the18 deaths (22% operative mortality) occur-red in this group. Patients in whom pro-longed conservative resuscitation had beenattempted fared worse (64% survival), asdid patients with poor hepatic reserve (Pughgrade C: 32% survival). Twenty twopatients (27%) rebled within 30 days, 18after urgent shunts, and 12 died. Seven(1 10/o) of the longterm survivors havesuffered recurrent variceal haemorrhage,with a clear relationship to shunt or portalsystem thrombosis. Portasystemic encepha-lopathy occurred in 13 survivors (20(%).We conclude that urgent lienorenal or

distal splenorenal shunts for unremitting orrelapsing variceal haemorrhage are worth-while. Selective decompression has beenadvocated as the precedure of choice, butthe non-selective lienorenal shunt is easierto perform and is equally acceptable interms of postoperative morbidity.

A medical/surgical team approach to themanagement of acute bleeding varices

D MUTIMER, P D WRIGHT, J FREFMAN, t ILOOSE,AND 0 F W JAMES (Departments of Medicine,Surgery and Radiology, Freeman Hospital,Newcastle Upon Tyne) While several con-trolled trials of different single modalities oftreatment for acute bleeding varices havebeen carried out it seems likely that atreatment plan allowing the use of more

A1386


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


than one method may prove more practical.We have evaluated the use of a 'multidisci-plinary' treatment algorithm in the manage-ment of such patients.The treatment schedule is centred on

early endoscopy (by physician or surgeonon a joint unit), with injection of varices atendoscopy, or insertion of Sengstaken tubewith subsequent endoscopic injectionwithin 24 hours. If bleeding persists despiteinitial injection (or if injection impossible)patients are submitted to gastric devascu-larisation and oesophageal transection(GDOT) or percutaneous transhepaticembolisation (PTE) the latter for poorprognosis (modified Childs C) patients.

Ninety two patients seen for first time infive years (1982-1986), had 131 emergencyadmissions with bleeding. Childs grade:27A, 22B, 43C (37 patients over age 60, 11under age 20.)One hundred and thirty one bleeds, one

death before treatment. Ninety one control-led with sclerotherapy±tamponade--84discharged, seven non-bleeding deaths.Thirty nine continued haemorrhage; 14 notreatment died, 14 GDOT, 11 PTE. Of 14GDOT bleeding was arrested in 12 (sevendischarged). Of 11 PTE bleeding wasarrested in nine (seven discharged).Thus of 131 admissions haemorrhage

controlled in 85%, discharge achieved in75%.We conclude this 'team approach' offers

results comparable with the best achieved in'single treatment' centres and in a relativelypoor risk group.

Histological improvement after antiviraltreatment for chronic hepatitis B virusinfection

M G BROOK, 1. PETROVIC, J A MCDONALD, P J

SCHEUER, AND H C THOMAS (AcademicDepartment of Medicine and Department ofHistopathology, Royal Free Hospital andMedical School, London) Sequential liverbiopsies taken from 50 patients with chronichepatitis B virus (HBV) infection under-going therapy with AraAMP or alpha inter-ferons were compared with biopsies from 25untreated controls. These were scored(1-4) for portal/periportal, lobular andoverall inflammation. In those clearing HBeAg and HBV DNA from their serum, therewas a significant reduction in portal/periportal (p<0-01), lobular (p<0)'01) andoverall (p<0-01) inflammatory activity ascompared with controls or non-responders,and none developed cirrhosis. In contrast,inflammatory activity continued in those

that did not respond to therapy or receivedno therapy, and two progressed to cirrhosis.Immuno-cytochemical studies confirmedthat in those clearing HBe Ag and HBVDNA from the serum, HBc Ag and HBe Agwere also lost from the liver.

Cirrhosis/fibrosis (p<0 10) and abscenceof orcein positive hepatocytes (p<0-05)were the only factors to predict response.The studies support the view that success-

ful treatment of chronic HBV infection isassociated with loss of hepatic as well asserum markers of HBV replication and isfollowed by a reduction in hepatic inflam-mation, preventing in some cases pro-gression to cirrhosis.

INFILAMMATORY BOWEL DISEASE POSIERS

Faecal short chain fatty acids (SFA) inulcerative colitis (UC) and controls

D A BURKE, P G R GODWIN, AND) 1 R AXON

(Gastroenterology Unit, the GeneralInfirmary, Leeds and Department of Micro-biology, University of Leeds, Leeds)Anaerobes form the bulk of the faecal floraand the SFA are an end product of anaero-bic metabolism. Little is known about theanaerobic flora in patients with UC. In vitrowork has shown a decreased ability tometabolise SFA by the colonic epithelium.To determine any disturbance in theequilibrium between production of meta-bolic products and their absorption andmetabolism by the colonic epithelium faecallevels of butyric acid (BA) and acetic acid(AA) were measured by head space gasliquid chromatography in 16 patients withUC in relapse and 20 controls and resultsexpressed as smol/g dry weight. In sevencases SFAs were assessed before and aftertreatment. Colitic faecal BA mean 149 andAA mean 438-5 Fmol/g were significantlyhigher than control BA, mean 28 ,tmol/gand AA mean 64-5 ftmol/g (p<0-005).There is a significant correlation betweenBA and AA levels in controls p<O.()05 butnot in colitics, mean faccal BA and AAbefore and after treatment were 126 imolBA/g, 187-5 tmol BA/g and 368-5 smolAA/g, 550-5 ,umol AA/g respectively (NS).This study shows that faecal SFA in patientswith UC are significantly raised and sug-gests either over production or decreasedutilisation of these fatty acids. These datasupport the in vitro studies showingimpaired metabolism of SFA in UC.

Inhibition of fatty acid oxidation as a causeof ulcerative colitis

W E W ROEI)I(iER ANt) S It NANtF. (INTRODI)U( I.I)

BY S C TRUEI OVE) (Department o'f SurgerYanid Cell Ph vsiology Laboratory, Univers'rityof Adelaide at Tlze Quleent ElizabethiHospital, Adelaide, A ustralia) Studies haveshown that ulcerative colitis is associatedwith inhibition of fatty acid oxidation incolonocytes and that inhibitors of fatty acidoxidation induced colitis in experimentalanimals. The hypothesis was tested whethera weak inhibitor of fatty acid oxidation(2-bromobutyrate, 4-mercaptobutyrate) inconjunction with a selective stimulator offatty acid oxidation (sodium nitrite) couldreproduce strong inhibition of fatty acidoxidation in colonocytes similar to the bio-chemical lesion observed in human colitis.Isolated colonocytes of the rat were used tomeasure fatty acid oxidation as "CO,production from l-'4CJ butyrate. 4-Mercaptobutyrate alone inhibited 14 5+2.40% and 2-bromobutyrate alone inhibited31.4+1 3°/0 of fatty acid oxidation. Com-bination of bromobutyrate and sodiumnitrite inhibited 73(0±8X5% of fatty acidoxidation. Bromobutyrate inhibited keto-genesis by 42 4±8+0(Y0 which was doubled(89.3+6±5'(Y) in the presence of sodiumnitrite. The proposed hypothesis Wasupheld from which it is suggested that aninhibitor of fatty acid oxidation from thecolonic lumen in conjunction with a stimu-lator of fatty acid oxidation, nitrite, knownto be generated from laminal cells arcneeded to producc the biochemical lcsionobserved in human ulccrative colitis.

Milk antibodies and circulating immunecomplexes in inflammatory bowel disease

J MAIN, tl MCKEN/IFI, C R PINNINGTON, ANt)t) PARRA IT (Department of Medical Micro-biology, Ninewells Hospital antd the Depart-ment of Medicine, Kings C(ross Hospital,Dundee) IgG and IgA antibodies againstwhole milk and certain milk antigens wereestimated by an ELISA in 44 patients withCrohn's disease, 42 patients with ulcerativecolitis and a control group (25). IgG aindIgA containing circulating immune com-plexes (CIC) were estimated by a con-glutinin binding ELISA. Anticasein IgAantibodies were significantly elevated inboth disease groups (p<0(05, Mann-Whitney). IgG and IgA containing CICwere significantly elevated (p<0.0)5, Mann-Whitney) in the Crohn's disease (IgGmean=374 tg/ml, range 0-3()00; IgA

A1387


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

TI-ue British Societx oJ Gastroenterologv

mean=47 ,ug/ml, range 0-1000) and ulcera-tive colitis (IgG mean= 192 [tg/ml, range

0-3000; IgA mean=42 .tg/ml, range 0-660)groups when compared with controls (IgGmean=83 ftg/ml, range 33-210; IgAmean=8 Rg/ml, range 0-75). There was a

significant correlation between milk anti-body and CIC values for IgG in Crohn'sdisease (r=0-78, p<0-01) but not in ulcera-tive colitis (r=-028, p>0.05). The correla-tion between milk antibody and CIC valuesfor IgA was significant for both Crohn'sdisease (r=0-82, p<0.0l) and ulcerativecolitis (r=0-54, p<0-0l).The observed correlation between CIC

and food antibody supports the possibleassociation between CIC and food antigenprocessing. In inflammatory bowel diseaseincreased absorption of food antigensmay generate high levels of CIC, althoughtheir pathogenetic significance remains

uncertain.

Colonic lymphocyte subpopulations ininflammatory bowel disease

M C ALIISON, 1. W POULTER, A P DHILLON, ANI)

R E POUNDER (Royal Free Hospital, London)Although lymphocytic infiltration is a well-known feature of ulcerative colitis (UC) andCrohn's colitis (CC), the location ofT and Bcell subtypes and their state of activationhave not been fully characterised immuno-histologically. Frozen sections of colonicbiopsies or resections from five patientswith active UC, 13 with Crohn's disease(five with colitis) and seven controls were

stained with a panel of monoclonal anti-bodies using an immunoperoxidase tech-nique. The antibodies were RFT Mix(recognising T cells), RFB Mix (B cells),Leu 3A (helper T cells), SN 130 (subsets ofT and B cells) and anti Tac (a marker of Tcell activation). The histologically normalsections and those from patients with CCdisplayed marker variations in mucosal Tcell populations that did not correlate withdisease activity. However enlarged basallymphoid aggregates with wide submucosalextension were found in all CC sections, andthese aggregates contained many Tacpositive cells. By contrast, marked mucosalinfiltration with all T cell subtypes, includ-ing many Tac positive cells, was found in theUC sections. B cells were present in

lymphoid aggregates but not in the mucosalinfiltrate.These observations offer evidence that

the immunopathogenesis of UC and CCboth involves a cell-mediated immune

response but show marked differences in

cellular localisation.

Macrophage and lymphold subpopulationsin the granulomata of Crohn's disease

Y R MAFIlDA. S PATEH, K WuJ, AND I) P JE-Wi-i I.

(G(astroentterology U,iit, RadclitfItnfirmary, OvJord) Granulomata arc a

characteristic feature of Crohn's disease.This study describes the distribution ofmacrophage and lymphocytc subpopula-tions in the granulomrata of four paticntswith Crohn's disease (thrce ileall, one

colonic).Tissue was snap frozcn and 4 u sections

stained using the peroxidasc and alkalinephosphatase techniqucs, singly or in com-

bination. Acid phosphatase (ACP) andnon-specific estcrase (NSE) stiaining was

also performed. Monoclonal antibodiesused: RFD1 (interdigitating cells), RFD9(epithelioid cells and tingible body macro-

phages), RFD7, EB I1 (tissue macro-

phages), 3C10, Y1/82A, UCHMI (mono-cytes and macrophages), 3G8 (Fcy receptoron polymorphs and some macrophages),CR3/43 (HLA-DR), BT3/4 (HLA-DQ),IL2R (interleukin 2 receptor), T310 (CD4),DK25 (CD8).The epithelioid cells and giant cclls were

strongly RFD1I+, RFD9+, EBII+,3C1()+, YI/82A+, CR3/43+, BT3/4+-,ACP+, NSE+, and moderately 3G8+,UCHMI+, and IL2R+. RFD7+ cells(weakly ACP+, NSE+) only occurred atthe periphery of the granulomata. Lympho-cytes within the granulomata were mainlyCD4+. CD8+ cells were few and scatteredmainly at the periphery in association withRFD7+ macrophages. This distribution is

in contrast with that seen in schistosomalintestinal granulomata.

Crohn's granulomatia comprise cpithe-lioid cells (RFD9+, RFD7- ) intimatelyassociated with CD4+ lymphocytes.Epithelioid cells are 11 2R+ and appear tobe activated.

Mucosal glycoprotein synthetase in Crohn'sdisease: a protective mechanism?

M C WINSIET, VAILERIEi POXON, A Al.l.AN, ANI)

M R B KEIGHili.Y (The Genieral Hospital,Birmin,gham) Glucosamine synthetiase(GS) is a rate limiting enzyme in mucosalglycoprotein synthesis. Increased rectal GSactivity has been reported in patients withileal Crohn's disease (ICD) which may

represent a protective responsc to toxicluminal factors. To test this hypothesis,rectal GS activity was measured in ICD(n= I 1), Crohn's proctitis (CFP n= 11) and a

control group (n= 11) before and two, six,

andL 12 weeks after a defunctioningileostomy. Fhe elffect of restorationi ofintestinal continuity was also assessed inCrohn's disease (CD, n=9) and controls(n=7). Predeifunction GS activity was sig-nificantly increased in the ICD group com-pared with controls (34X8±4-5; 22 9±3+4,p<0-05). Postdefunction - there was nochange in GS activity in the control groupbut ICD activity had fallen significantlyfronm 34-8±4-5 to 19-2±3-4 at six weeks,and 16-4±4-6 at 12 weeks, p<0()-. Crohn'sproctitis activity had fallen significantlyfrom 28X6+3(0 to 16-2+2-5 at six weeks,p<0(l0. On restoration of intestinal con-tinuity, there was a significant increase inCD from 14 7+2-5 pre-operatively to25-5±3-7 at two weeks, p<0-0l but nochange in the control group. These resultssuggest that a luminal factor significantlystimulates glycoprotein synithesis in Crohn'sdiscase. This finding supports the role of atoxic faecal component stream in the patho-genesis of Crohn's disease.

Personality in inflammatory bowel disease(IBD): primary, or secondary to chronicillness?

I) A F ROBERI'SON, J RAY, AN[) I DIAMONI)

(Departments ofJ Medicine atnd Social Statis-tics, UniversitY of Soutlihampton, Southamp-toni) The high prevalence of personalitytraits of neuroticism. anxiety and intro-version is well described in IBD, butwhethcr these arc secondary to disease, or aprimary phenomenon, is unknown. Wehavc measured the Hospital Anxiety andDepression scale (HAD) and Eysenckpersonality inventory (EPI) measuringneuroticism and extraversion on a linearscale (1-24) in 8() IBD patients, in 22 newlyreferred patients before the diagnosis ofIBD, and in 4(0 controls (Diabetics attend-ing a clinic). Inflammatory bowel diseasepatients were more neurotic (mean EPI12-(0 9, p<0.05) and less extroverted (9 8 v12-0, p<0-05) than controls and these traitsare prcsent before diagnosis (EPI neuro-ticism 13 v 12, EPI extraversion 9-8 v 9-8,NS). In established disease, depressioncorrelates strongly with disease activity andintrovcrsion with disease duration. Patientswere aware of a close link between stressand disease activity: 59 identified a stressfullife event which they believed precipitatedthe illness. Neuroticism and introversionare common in IBD patients and these traitsarc present before diagnosis. Depressionoccurs predominantly in those with activedisease, and introversion increases as thedisease progresses.

A 1 388


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Smoking and sugar intake are separate butinteractive risk factors in Crohn's disease

B KATSCIIINSKI, R F A l.OGAN, M E)MOND, ANt)M J S LANGMAN (Departments of' Thera-peutics anid Communit AIMedictine aIltdEpidemiology UniversitsV of' Nottingham,Nottingham) Previous studies have consist-ently found a strong positive associationbetween sugar intake and Crohn's disease(CD) and recently others have found astrong association between smoking andCD. As sugar intake and smoking are oftenlinked we have investigated this relation-ship in 104 CD patients and 153 matchedcommunity controls. Data on added sugarintake (AS), confectionery consumptionand smoking habits were obtained by postalquestionnaire. Relative risk (RR) estimates(95%o confidence limits) were derived usingMantel-Haenszel analysis. Added sugarintake and smoking were associated withone another. After adjusting for AS smok-ing still showed a significant association withCD with a RR of 1-8 (1-04-3-2). Equallyafter adjusting for smoking AS was alsostrongly associated with CD in a dose-response pattern (x5=7-7; p<-*025) but thisrelationship was only evident in never andex-smokers. For never and ex-smokers theRR's associated with no AS, <50 g/day and>50 g/day were respectively 10, 1*8 (0X-83-0) and 4.6 (2-0-11) (X =12-1; p<0-005)and for smokers 1-0, 0(8 (0-2-2.7), 1-1(0.4-3 1) respectively. The AS relationshipwas supported by a separate associationwith confectionery consumption with theRR's of never or < monthly consumption,monthly, and weekly or more often beingrespectively 1-0, 1-7 (0-7-4-0) and 3-0 (1-2-7-5) (X2 =6-8; p<0.05). These findingsindicate that while smoking and AS areindividually associated with CD combinedexposure results in no further increase inrisk, suggesting that they may operatethrough a common mechanism.

Are clinical activity indices helpful in assess-ing active intestinal inflammation in Crohn'sdisease?

G 1' CRAMA-BOHIBOUTH, I BIEMOND, A S PENA,If W VERSPAGI:'r, I I WETERMAN, AN[) C B H W

LAME.RS (Dept Gastroenterology andHepatology, University Hospital Leiden,The Netherlands) Quantification of thefaccal excretion (FE) of iv administeredautologous Indium- 111 tropolonate label-led granulocytes is a reliable parameter ofbowel inflammation. In this study we corre-lated 24 h Indium-l 11 FE with several

clinical activity indices for Crohn's disease(CD) such as CDAI (Best et al). Al (vanHees), SI (Harvey and Bradshaw), OxfordScore, and laboratory parameters: ESR,serum albumin, orosomucoid, C-reactiveprotein, a,-antitrypsin (at -AT) faecal con-centration, and ca1-AT clearance. Twentyeight CD patients (18 F, age 16-60 years)were studied. Nine had colonic involve-ment, 18 had small bowel localisation, onepatient colonic and small bowel involve-ment. Indium-llI activity of faeces sampleswere measured in a gamma counter andexpressed as per cent of the injected dose.The median was 8% (range 0-.15-50%) forcolonic involvement and 0)*75% (range 0(03-2 10%) for small bowel disease.No significant correlation was found

between Indium- I ll FE and any of the fouractivity indices used in this study, neitherfor colonic nor for small intestine or for bothcombined. Indium-11 FE however waspositively correlated with orosomucoid(r=0-53; p<0-05), al-AT faecal concentra-tion (r=0-64; p<0-05), a,-AT clearance(r=-039; p<0-05), and negatively corre-lated with serum albumin (r= -0-52;p<0)05).None of the studied activity indices reflect

accurately the degree of inflammation asassessed by Indium-11 FE. Objectivelaboratory parameters should be used instudying activity of CD. Indium-111 FE istherefore a good parameter which can beused in the evaluation of new forms oftreatment.

Observer variation in the histopathologicalassessment of rectal and colonic biopsies

A 'FHEODOSSI, D J SPIEGELHALTER, J JASS,J FIRTH, M DIXON, M IlEADER, D LEVISON,R lINDLEY, I FilIPE, A PRICE, N A SHEPHERD,S THOMAS, AND H 'I'HOMPSON (The MaydayHospital, MRC Biostatisticy Unit Cam-bridge, Departments of Pathology, StMarks, Westminster, St Batholomews,Guys, Northwick Park Hospitals, LeedsGeneral Infirmary, and The GeneralHospital, Birmingham) If skilled histo-pathologists disagree over the interpreta-tion of the same biopsy this leads to diagnos-tic error. The aim of the present study wasto determine the magnitude of variationamongst ten observers with a specialinterest in gastrointestinal histopathology,who independently interpreted the samebiopsies for morphological features whichmay discriminate between patients withCrohn's disease (n=24), ulcerative colitis(34), and normal subjects (18). More than50 000 items of information were evaluated.

Although 38 of 51 features had agree-ment measures significantly better thanexpected by chance, only three (neutrophilinfiltration of crypt epithelium, neutrophilsin lamina propria and Langhans giant cells)had Kappa values of 0-4 or grefater, suchvalues being considered to show fair to goodagreement. Thus, the vast majority offeatures had Kappa values indicating pooragreement beyond chance. The range ofagreement amongst the 45 observer pairsover the final diagnosis was 65-76%). Onlynormal biopsies had a Kappa value greaterthan 0-4. Those from patients with ulcera-tive colitis and Crohn's had Kappa values of0(3 and 0)2 respectively.These results suggest that there is con-

siderable room for improvement in thereliability of colonic biopsy interpretationand could probably be achieved using moreexact definitions of morphological featuresand diseases.

Tc-99m-HMPAO labelled white cell imaging- a new technique for the dynamic assess-ment of inflammatory bowel disease

A H M tlEAGERTY, D C COSTA, D UlI, J M -CLARKE, S R CAIRNS, P J EL.l, AND N I MCNEIIt(Department of Gastroenterology andInstitute of Nuclear Medicine, MiddlesexHospital, London) The assessment ofactivity and extent of inflammatory boweldisease (IBD) by endoscopy or radiology istime-consuming, uncomfortable, oftentechnically difficult and open to misinterp-retation. Using tracer labelled white cellsfor dynamic imaging to assess the extent ofinflammation has much improved the situa-tion.Tc-99m-Hexamethylpropyleneamincox-

ime (Tc-99m-HMPAO) has been identifiedas a useful tracer for white cells. This wassimultaneously compared with the estab-lished Indium-IlI -oxine (In-lI1) labellingtechnique in 12 patients with known orsuspected IBD at one, four, and 20) h usingappropriate energy windows for the iso-topes. One patient had scans seven daysbefore, and 60 days after resection of ilealCrohn's disease as well as imaging of theresected specimen.

This study showed that both isotopesproduce identical leucocyte distributions inthese patients, ranging from nothing toextensive intestinal uptake in small bowelCrohn's disease. The resolution with Tc-99m-HMPAO was uniformly better thanwith In-1il. Scans of the operative speci-men showed localisation specific to theactive disease.

A 1389


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Assessing the extent of IBD with Tc-99m-HMPAO labelled white cells has the princi-pal advantage of better resolution withoutmissing active disease and utilising a morereadily available tracer. We believe it is thestandard isotopic method of choice inassessing IBD, until an in vivo labellingtechnique is widely available.

Effects of prednisolone, sulphasalazine,5-aminosalicylic acid and indomethacinon prostaglandin E2 and leucotriene B4production

J J KEATING, W J MAXWELL, F P HOGAN, AND

P W N KEELING (Dept of Clinical Medicine,Sir Patrick Dun Research Laboratory, StJames' Hospital, Dublin) Sulphasalazine(S/S) and 5-aminosalicylic acid (5-ASA)decreases the frequency of relapse, whilstprednisolone (Pred) controls the acuteinflammatory response in patients withinflammatory bowel disease (IBD). Theaims of our study were to compare the invitro effects of S/S, 5-ASA, Pred and indo-methacin (INDO) on eicosanoid [prosta-glandin E, (PGE2) and leucotriene B4(LTB4)] biosynthesis by stimulated peri-pheral blood mononuclear cells frompatients with IBD. Cells were stimulatedwith opsonised zymosan and PGE, andLTB4 biosynthesis assessed by radio-immunoassay. Indomethacin and Predinhibited PGE2 production (2-3+0.2, n=6 v8-0±1 8, n=7 v 15-4+3, n=7, p<0-05 ngPGE,/106 monocytes, INDO v Pred v con-trol). Neither drug significantly alteredLTB4 production. In contrast S/S and5-ASA increased PGE, concentration(44-6+11-6, n=7, 48-5+10-8, n=7 v15-4+3, n=7, p<0.025, cells cultured withS/S v 5-ASA v control cells). Again neitherdrug significantly altered LTB4 production.The results suggest that pharmacologicalmanipulation of PGE2 and LTB4 biosyn-thesis may play a different role in themanagement of acute and chronic IBDhence Pred decreases PGE, productionduring an acute episode and may decreasethe inflammatory reaction whilst S/S and5-ASA increase its synthesis and hencemodulate the chronic immune response.

Oral tobramycin improves the outcome ofacute ulcerative colitis (UC)

D A BURKE, S A CLAYDEN, M F DIXON, A T R

AXON, AND R W LACEY (GastroenterologyUnit, The General Infirmary, Leeds andDepartment of Microbiology and Depart-

ment of Pathology, University of Leeds,Leeds) Patients with UC harbour adhesiveE coli in their colon. This double blind trialwas performed to determine whether eradi-cation of E coli is of benefit. Eighty fourpatients with UC in relapse were random-ised to seven days tobramycin (T) 120 mgtds or placebo (P) as an adjunct to steroidtreatment. Patients were assessed byclinical and histological scoring at entry andendpoint (EP) (21 days for inpatients, 28days for outpatients). Culture for faecal Ecoli was undertaken at entry, seven daysand EP. Tobramycin and P groups werewell matched for age, sex, extent of disease,length of history and duration of relapse.Bacteriology results were unsuitable foranalysis in 6 T and 7 P. In the remainderseven day assessment showed that E colihad been eradicated in 77% T and 8-8% P,p<0-0)01 and the strain changed by EP in91% T and 13-8% P, p<OO()Ol. From eachgroup one patient went to surgery and onefailed to attend during the trial. Clinical andhistological scores in each group weresimilar at entry but at EP were significantlylower in the T group (p=0-04 and 0(03respectively). Complete symptomaticremission was obtained in 31/42 of T and18/42 P, p=O)008. Ten of 10 first attacks inT and 6/10 in P achieved remission by EP(p=0.04). Oral tobramycin successfullyeradicates the original E coli strain in atleast 77% of patients with UC and is associ-ated with a significant clinical and histo-logical benefit.

A comparison of delayed-release 5-animo-salicylic acid (5-ASA) and sulphasalazine(SSZ) as maintenance treatment of ulcera-tive colitis (UC)

S A RILEY, V MANI, M J GOODMAN, AND A

TURNBERG (University Dept of Medicine,Hope Hospital, Salford, Leigh Infirmary,and Bury General Hospital, Manchester)Many patients with UC are denied thebeneficial effects of SSZ as side effects arecommon. As most of these are related to thesulphapyridine component of the drug wehave assessed the efficacy of a delayed-release 5-ASA preparation in maintainingUC remission.One hundred patients (51 M:49 F, aged

20-78 years) with quiescent UC wererandomly allocated to either enteric coatedSSZ or equivalent dose 5-ASA in a double-blind 48 week trial. Eight patients werewithdrawn (four non-attendance, two non-compliance, one stomatitis (SSZ), and oneinappropriate inclusion) leaving 44 SSZ and

48 5-ASA for analysis. Groups were com-parable for age, sex, duration, and extent ofdisease.

Relapse rates at 48 weeks were SSZ38-6% (95% confidence limits 24-54%) and5-ASA 37.5% (95% confidence limits 24-53%) X2=0-01, p>0-90. Mean time torelapse and relapse severity were similar inthe two groups. Headaches and uppergastrointestinal symptoms were lesscommon in the 5-ASA than in the SSZgroup. Haemoglobin concentration roseand MCV fell during 5-ASA treatment,folate levels, however, remainedunchanged.

Delayed release 5-ASA is as effective asSSZ in maintaining UC remission and hasfewer side effects.

Combination antimycobacterial chemo-therapy with major clinical response inestablished Crohn's disease

M PARKER, S HAMPSON, S SAVERYMU1-rU, SCAWTHORN, J J MCFADDEN, J THOMPSON, E

GREEN, K RUrrER, AND J HERMON-TAYIOR(Department of Surgery, St George'sHospital Medical School, Cranmer Terrace,London) Some Crohn's disease (CD)mycobacterial isolates have been shown tobe indistinguishable from M paratuberculo-sis on the basis of restriction fragmentlength polymorphisms using cloned myco-bacterial DNA probes. Seventeen patientswith established Crohn's disease weretreated with rifampicin, isoniazid, etham-butol and pyrazinamide for up to ninemonths. Disease was assessed by CDactivity idex (CDAI) monthly, and byIndium scanning, colonoscopy, histologyand radiology before, three and ninemonths after treatment. Three patientshave not yet completed initial assessment.Treatment failed in two patients because ofnon-compliance although one of these hadresponded before withdrawal. Eleven of 12patients responded, mean initial CDAI of275±99-7 (ISD) falling to 176+93 withinone month. In five patients with severeintractable CD refactory to conventionaltherapy, response was dramatic andimpending surgery was avoided. Althoughdisease course after cessation of chemo-therapy is not yet known, these early resultswould support a wider study of thisapproach in the clinical management of CD.The consistency, and in some cases, scale ofthe response further suggests that patientswith otherwise uncomplicated severe CDshould not be subjected to major surgicalresection before a trial of antimycobacterialchemotherapy with multiple drugs.

A1390


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Fish oil for inflammatory bowel disease?

P J PRICHARD, T K DANESHMEND, J E WILLIARS,B FILOPOWICZ, D L[ESER-SMITH, AND C JHAWKEY (Departments of Therapeutics,University Hospital, Nottingham) Leuco-trienes synthesised from the precursoreicosapentaenoic acid (EPA) - that is,LTB5, appear to be less pro-inflammatorythan those derived from arachidonic acid -that is, LTB4. Administration of fish or fishoil high in EPA may thus be beneficial ininflammatory conditions such as ulcerativecolitis. We have investigated the bio-chemical consequences, safety and efficacyof this dietary approach. Consumption offish oil (Hi-EPA, 20 ml/day, giving 4 gramsof EPA per day) for one month caused anincrease in cell membrane EPA from1-17±0-16% (mean+SE) to 3-35+0 34% oftotal fatty acids (GC assay), a 55% (95%confidence limits: 40-70%) decrease in thedetectable amount of the pro-inflammatorycompound LTB4 (HPLC assay). Produc-tion of the less inflammatory compoundLTIB5 was detectable in only four of 12subjects under control conditions, but waspresent in eight subjects after fish oil. In anopen pilot study, five subjects with severecolitis resistant to steroids were treated withthe addition of Hi-EPA, 20 ml/day, andazathioprine (2 mg/kg/day) to existingtherapy. After three to four weeks, all hadimproved and four were in remission. Fishoil caused no significant adverse effects andits use in inflammatory bowel diseasewarrants further investigation.

Controlled trial of clofazimine in Crohn'sdisease

N H AFDHAL, A LONG, J LENNON, J CROWE,AND D P O'DONOGHUE (Dept of Gastroenter-ology, St Vincent's and Mater MisericordiaeHospitals, University College, Dublin,Ireland) The aetiology of Crohn's diseaseremains obscure and an infective agent,perhaps an atypical mycobacterium, hasbeen proposed. A preliminary study withclofazimine (CL), a broad spectrum anti-mycobacterial agent with both anti-inflammatory and neutrophil stimulatingproperties, was favourable but hasremained unconfirmed (Gut 1982; 23: 449).Therefore to determine the effect of CL oninducing disease remission, 49 patients (34F, age 28±13 years) with active Crohn'sdisease were randomised to receive cortico-steroids and either CL 100 mg (25) ormatching placebo (PL, 24). Disease activitywas monitored using a modified disease

activity score (DAS) combining subjectivesymptoms with changes in haemoglobin,ESR and serum albumin. Remission wasdefined as reduction in DAS-4, and steroidwithdrawal within three months with noevidence of relapse for a further month. Atrandomisation (induction phase) bothgroups were similar for site of disease (largebowel 25, small bowel 24), number ofprevious relapses, median DAS (CL 10, PL9-5) and median initial steroid dosage (45mg). Twenty seven patients (55%) achievedremission (median DAS: 1) and there wasno difference between the two groups (CL15: PL 12, p=NS). In the second phase ofthe study (maintenance phase) these 27patients continued to receive either CL orplacebo for a further 8 months or untilrelapse, defined as an increase in DAS-8.Eighteen patients (38%) had no furtherrelapse (CL 12: PL 6, p=NS) and success-fully completed the trial. Side effects wereminor and consisted of skin rash (CL 3) andincreased pigmentation (CL 12, PL 5).

In conclusion, clofazimine was neithereffective in inducing or maintaining remis-sion and is further indirect evidence againsta mycobacterial aetiology for Crohn'sdisease.

An audit of the management of Crohn'scolitis

H A ANDREWS, P LEWIS, J ALEXANDER-

WILLIAMS, M R B KEIGHLEY, AND R N ALLAN

(Gastroenterology Unit, The GeneralHospital, Birmingham) An audit of the longterm outcome among 361 patients withcolonic Crohn's disease followed between1944-1986 (mean follow-up 13-3 years) hasbeen undertaken in an attempt to defineoptimal management.

Right sided disease+ Distal ileum (n= 85):11% were treated conservatively. Theremainder underwent resection (meaninterval from diagnosis 3-8 years). Cumula-tive re-operation rates for recurrentdisease (cum-re-op) were 19%, 40%, 51%at 5, 10, and 15 years. Eighty per cent ofsurvivors are symptom free although 29%have radiological evidence of recurrentdisease. Only three patients are currentlyunwell.

Left sided disease (n=131): 37% weretreated conservatively. Of those under-going resection (mean interval six years) thecum-re-op were 39%, 50%, and 55% at 5,10, and 15 years. Ninety five per cent ofsurvivors are currently well. Fifty five percent have a permanent stoma and 45%radiological evidence of disease.

Extensive colonic disease (n= 145): 22%were treated conservatively. Of thoseundergoing resection (mean interval 6.7years) the cum-re-op was 48%, 61%, and79% at 5, 10, and 15 years. Currently 92%of survivors are well although 45 (41%)have evidence of recurrent disease and 78(67%/) have a permanent stoma. Only ninepatients are symptomatic.

In total there were 36 related and 24unrelated deaths. The commonest causes ofrelated death were sepsis (16), electrolyteimbalance (seven), and cancer (six).The longterm prognosis is good even

though the incidence of reoperation forrecurrent disease is high. Most patients arecurrently well although many have to accepta permanent stoma to achieve this status.

Segmental colectomy for Crohn's disease ofthe colon

H A ANDREWS, A ALLAN, C J HILTON, M R B

KEIGHLEY, R N ALLAN, AND J ALEXANDER-WILLIAMS (Gastroenterology Unit, GeneralHospital, Steelhouse Lane, Birmingham) Itis now clear that limited resection is appro-priate for small bowel Crohn's disease par-ticularly as the risk of re-operation is notaffected by histological evidence of involve-ment at the resection margins. This studyevaluates the place of limited resection incolonic disease.

Thirty six patients underwent segmentalcolonic resection for Crohn's colitisbetween 1944 and 1986. None of the 29patients undergoing segmental resectionand primary anastomosis had anastomoticdehiscence. There were two postoperativedeaths from septicaemia in elderly patients.At 10 years the cumulative re-operation

rate was 60% compared with 48% aftercolectomy and ileorectal anastomosis, and24% after panproctocolectomy. There wasa significantly higher recurrence rate in thesegmental colectomy group (X' p<0-(X)8)compared with the recurrence free intervalcurves following total colectomy withileorectal anastomosis or panproctocolec-tomy.There was no apparent functional or

nutritional advantage from retainingresidual functioning colon following localcolonic resection. We now rarely recom-mend segmental colectomy for colonicCrohn's disease.

Sustained epithelial proliferation afterfaecal diversion for Crohn's disease objec-tive evidence of mucosal recovery

A 1 39 1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society ofGaistroenterology

M C WINSIL[I, [)ENISE YOUNGS, A ALtLAN, AND)

M R B KEHitLEUY (The General Hospital,Birmingham) Faecal diversion for Crohn'scolitis frequently produccs clinical remis-

sion but there is no objective evidencc of

mucosal recovery. The effect of faccaldiversion may be mcdiated by changes in

mucosal cellular kinetics. To test this hypo-thesis, the cell birth rate (crypt ccll produc-tion rate - CCPR) ol rectal mucosai in

patients with ileall Crohn's diseaisc (ICD,n= 1), Crohn's proctitis (CP, n=9) and a

control group (n= 10) was mceisured beforcztnd two, six, and 12 weeks after defunc-tioning ileostomy. Crypt cell productionrate was assessed by an in vitro stathmo-kinetic method. Thcre was no significantdifference in pre-diversion levels of CCPR(control =34±10(), ICD =28±()06, CP=2.4+).9). Post-diversion therc was a signifi-cant reduction in the CCPR of the controlgroup at two weeks (I12±0)4, p<0()05) butnot at six weeks (1 7±0()4) or ait 12 wceks(3.3+ 1.5). There was no significant changein CCPR of the ICD or CP group at twoweeks (ICD=20()±0(7; CP=3±()8(),8) sixweeks (ICD=2-1_±0(8; CP=38+ 1 -)) or 12

weeks (ICD=2 1±0)5; CP=2 1±16). Thesustained proliferation rate after diversionin Crohn's disease compaired with the sig-

nificant fall in the control group may repre-sent a regenerative mechanism promotiingepithelial repair. This response suggeststhat a toxic faecal component may per-petuate the inflammatory process in

Crohn's disease.

Enterovaginal fistulae in Crohn's disease

FRANCOISE I1EYEN, M WINStLET, 11 ANDREWS,R N AllAN, M R B KUIGHIU.Y, AN) J

AlrEXAND)UR-WiILLIAMS (The Gen-ieralHospital, Birmingham) Conservativemanagement of enterovaginal fistulae in

Crohn's disease has been advocated for lesssymptomatic patients. We reviewed 27patients with spontaneous enterovaginalfistulae complicating active disease, seen

between 1970 and 1987. Ten required earlyoperaition, including three proctectomies.Conservative management healed no listulain 16 patients. During this long follow up(two months-35 years), six patientsrequired a panproctocolectomy. one

patient had a total colectomy for megacolonand one had a defunctioning stoma forstricture. Two symptomatic patients died ofunrelated causes, four are still mildly symp-tomatic. Two patients died of malignancyarising from within the chronic entero-vaginal fistula. Although some entero-

vaginial fistulae causc little local disahilityand cain be mainaged symptomatically hvscrupulous hygiene, they do not heal on

conservative measures and in time themiajority develop such severe bowel symp-toms that resection, usualily proctectomy. is

required. As two patients in this groupdeveloped cancer in the fistula track, we

would now recommend early considerationof radical surgical thcrapy for cntcrovaginalfistulac in Crohn's diseasc.

Studies on the anorectal function in ulcera-tive colitis (UC)

S S ( RAO, N W Ri At), ANt) C I) 11O DSWORITit

(Sub-Del)e (irtnen oJ F/ontionla Glastro-intiestina(il Pl/iysiOlogy5 a(1( N Uitritiotn, Roval

Hallnmshire Hospitll, She ffield) The patho-physiology of the howel disturbance if UCwats assessed by measuring anorectllpressures at multipie sites under basal coIn-ditions, during hbilloon distension of therectum aind during rectal infusion of salinein 29 patients with UC and in 30( norimllsublects. Resting and squeeze sphincterpressures were similar in the three groups.

[Ihe lowest rectal volume that could be

perceived, the volume required to induce a

desire to defaccate and the maximum toler-able rectal volume were all lower in alctivecolitics thian quiescent colitics (p<() ()()I)and controls (p<0(00 I). The rectal volumerequired to cause a sustained anal relaxa-tion was lower in active colitics (p<()0()5)than controls. During balloon distensioniboth initial aind steady state rectal pressures

were higher in patients with active thainquiescent colitis or controls (p<0(05).During resting conditions, rectail contrac-tions were recorded less frequently in activecolitics (12 5'Y(0) than in quiesccnt colitics(31 5'0o) and controls (37(%). During salineiniusion, the amplitude of rectal contrac-tions wals higher in colitics compared withcontrols and the volume of silinc retainedwats lower (p<0(001 ). In conclusiol, thefrequent and urgent defaecation in aIctivecolitis is relatted to a hypersensitive rectuimithat is morc reactive to distension andinduccs .a more pronounced sphincterrelaxation.

Detection of asymptomatic inflammatorybowel disease while screening for colorectalcancer

K ( BAIIANIYNI., .1 I MAYBIJRY, (G PYIF,

( MAN(iIAM, ANI) J I) IiARDCASTITI (Depaort-itietit of Sutgery,% UniversitY Hosifital,

Nottinghaflm) Previous studies based on

hospital diagnosis of symptomatic patientshatvc estimatcd the prevalenice of inflammia-tory bowcl diseaise (IBD) to be betwecn

) In Nottinghalm the exist-ence of a screening programme for color-ectal cancer hats provided an opportunity forthe detcction of apparently asymptomaticcases of IBI) and the prevalence to heassessed.Between 1983 and 1987, 37000 indi-

iduals aged 5(t-75 years have been offeredfaecal occult blood tests (FOBT), eitherthree or six davy Hilemoccult or Feca EIA, as

a screening test for colorectal neoplasia.Seventeen thousaind nine hundred aindthirty individuals have completed theFOBT's and of these 481 have been posi-

tive. Colonic investigtation in these indi-viduals revealed eight patients with pre-

viously undiagnosed IBD; five patients hadtotal ulcerative colitis, one proctitis, andtwo Crohn's disease. Two further patientswith ulcerative colitis were identified whohad been lost to follow up for 25 and 45years respectively. The prevalence ofulcerative colitis in this population was atleast 446 /100)00(95)(o Cl II 2-78 1) whilethat of Crohn's disease wats 11 2/100(0)00(95%, Cl (0-446). Thus the combined pre-

valence of IBD wias 56/100 000) and this datasuggests that current epidemiologicalstudies underestimate the UK prevalence ofIBD by at least 30%o.

B\ASKI SCI1NCt

Coeliac sprue (CS): the sequence ofimmuno-pathological events after rectal glutenchallenge

1) E ILOFT, P T CROWIF ANU) M N MARSH

(D)epartment of Medicinle, Hope Hospital,Solfort, Manchester) To document thechanges in rectal mucosal structure, epithe-fial lIymphocyte populattions and mast cellstollowing local gluten challenge in CSpatients and controls (C).

Biopsies were taken pre-, and at 1, 2, 4, 6.8, 12, 24. 36, 48, 60, 72, 96 hours post-challenge, with either 2 g FF3 (n=9 CS/6 C)or 500 mg ji lactoglobulin (n=6 CS/6 C)instilled rectally. Each biopsy was quanti-fied by computerised image analysis for: (1)lamina propria volume (LPV), surface, andcrypt epithelial volumes; (2) absolute cryptepithelial lymphocyte population (CEL).(3) Epitheliil lymphocytes were pheno-typed with monoclonal antibodies, (4)

A 1 392


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


vascular permeability monitored withperoxidase-anti-fibrinogen antibody and (5)mast cell structure assessed by EM.There was a marked rise in OKT8+

suppressor-cytolytic CEL maximal at sixhours (115% increase) in CS (p<0)-01).Lamina propria volume increased at onehour and significantly so by six hours(p<t)-t)5) reflecting mucosal oedema due tomicrovascular extravasation, as shown byextensive fibrinogen deposition within thelamina propria one hour postchallenge.There was no EM evidence of mast celldischarge. There wals no response to jilactoglobulin in CS or C.

In CS, rectal gluten instillation: ( 1) causesa rise in OKT8+ CEL maximal at six hours;(2) evokes an early increase in micro-vascular permeability and tissue swellingthat is not mast cell mediated or due to thelater lymphocytic infiltrate.

Synergism of hepatotropin, a novel hepato-trophic factor, and insulin on DNA in rathepatocytes

C SELTDEN AND) H J F HODGSON (D)ept ofMedicine, Royal Postgraduate MedicalSchool, Hatnnersmith Hospital, London)Liver regeneration can bc modulated byglucagon, insulin and epidermal growthfactor (EGF). We report the in vitrohepatotrophic activity of a high molecularweight fraction of human and rat serumafter partial hepatectomy, and identifiedthe factor in rats as a protein of subunitmolecular weight approximately I()()()(daltons on SDS-PAGE. In this study wetested hepatotropin in hepatocyte cultureswith glucagon, insulin and EGF. There wasa purely additive effect on DNA synthesiswhen hepatotropin was added to glucagonand EGF, but a striking synergistic effect onDNA synthesis when hepatotropin wasadded with insulin. 4"1-Thymidine incorpo-ration in the presence of insulin alone was21 100, 16840, and 11 740 dpm/20(000 cells(10 ", 10 ' and 10" M insulin); hepatotropinwas 28632 dpm/20)0)( cells. Thymidineincorporation at the same concentration ofhepatotropin with insulin was 58552,61 428, and 59t)0X dpm/20 000 cells at thethree insulin concentrations. (Quadrupli-cate cultures in three separate experi-ments) .

This striking synergism between insulinand hepatotropin requires further investiga-tion. Possibilities include positive co-operativity between the membranereceptors, as occurs between insulin andsomatomedin, or the activation of diffcrentintracellular second-messengers.

Different effects of VIP and PHI on intestinaladenylate

J A SMITIH AND R (G l ON(G (Medical ResearchCentre. Cit'v Hospital, Noltiglhatn) Vasoac-tive intestinal peptide (VIP) and peptidehistidine isoleucine (PHI) are closelyrelated regulatory peptides. Vasoactiveintestinal peptide is an intestinal secre-tagogue that exerts its effect by increasingintracellular cAMP levels. The adenylatecyclase of intestinal epithelial cells appearsto have receptors specificilly for VIP butnot for PHI. Peptide histidine isoleucine hasbeen reported to have an affinity for VIPreceptors. Fluoride ions are used to stimu-Iite adenylate cyclase ictivity in a non-specifc way via guanine nucicotide bindini,N proteins.The effect of VIP and PHI on basal and

fluoride stimulated adenylate cyclase wasassessed in homogenaltcs of normal humanduodenal biopsies. Results are expressed asmean+SEM pmol cAMP/minute/mg pro-tein. VIP ( 1) ' M to 10 ' M) and PHI ( I)M to 1() M) stimulated hasal adenylatecyclase. Maximal stimulation of basalladenylate cyclase (10)5±+322) was achievedwith VIP (19-55+7-39) and PHI (22-28+6-43) at a concentration of t) 'M. Both VI Pand PHI werc inhibitory at 10 M. In thepresence of 10 mM sodium fluoride PHIfailed to stimulate further whereaIs VIPcontinued to stimulate adenylate cyclase ina linear manner. Thus in vitro PHI maystimulate adenylate cycliase in a diffcrentway to VIP.

Endocrine tumours in ARG-vasopressin/simian virus 40 large T antigen transgenicmice: a model for human multiple endocrineneoplasia type I syndrome?

A k BISH(OP, 0 RINDI, I) MURPHY, J HANSON,G W 11 STAMP, M MURPHY, B HO(AN, AND J MPOTLAK (Dept.s oJ Histochemnistry anid Histo-pathology, RPMS, Hatntnerstni th Hospital,Dl) Cane Rd, Lonidon, Unit of VeterinaryCellular and Molecular Biology, RoyvalVeterinarY(C'llege, London acnd Labora-toryi of'Molecular Biology, NIMR, Lonidoni)A construct comprising the putative regulat-tory sequences of arg-vasopressin and theSimian virus 40 (SV40)) regions coding forlarge T-antigen was prepared and insertedinto the mouse genome in order to study invivo the effects of the expression of theSV40 oncogene. The offspring of these micedeveloped neoplisms of the pituitary andpancreatic fi-cells. Immunocytochemistry

was used to characterise these tumours. Thepituitary growths lfiled to show immuno-reactivity to antisera against pituitaryhormoness vaisopressin or at wide rainge ofother hormonall peptides and generalneuroendocrine mairkers. The endocrinetumours ol the paincrcas showed immuno-reiactivity lor insulin with occasionalscattered cells containing other pancreatichormones, particularly pancreatic poly-peptide. At the ultrastructural level, theendocrine niature of the pituitiary lcsions wasdefined by their content of granules, albeitsparsely distributed, and the pancreatictumours haid well-developed ji granules. Inboth lesions, SV40) large T-antigen could bcimmunostaiined in the nuclei and shown, byelectron microscopy. to occur in the hetero-chromattin. If these tumours aIre a rcsult ofthe inheritance and expression of SV40), thismay prove to be a useful model lor thehuman multiple endocrine ncoplasil Type Isyndromc.

Detection of the acetaldehyde adduct onhepatocyte membranes of rats fed ethanol

A J K WIll.IAMS, R t- BARRY, K YOUN(G, ANI) M

IIOPTON (Departtnett oJ Medicine, BristolRoval lnlfirtnaryt, Bristol) We hiave pre-viously shown thait acetaldehyde can bind(idduct) to amino aicid residucs in hepato-cyte membrane proteins and this cianactivate complement aind stimulate neutro-phil degratnulation aind superoxide release.We haive therefore determined the

presence of acetaldehyde adducts in vivo onthe hepatocyte membranes of rats fed 10(v/v) ethainol for 12 months+disulfiram ( 100mg/kg (o) daily). Liver membranes wereprepared by rapid percoll centrifugation,ind the amino acid profiles iater acidhydrolysis were determined by a photo-metric ninhydrin method. The position ofthe acetaildehyde-lysine adduct wais deter-mined by incubating poly-lysine withacetaldehyde, and this adduct wais detectedon the membranes of 23% (three of 13 rats)fed ethanol, and in 40'0 (two of five) fedethanol and disulfiraim but not in controlrats (not fed ethanol). The mean adductpeak sizes for both groups were similalr. Noacetaldehyde idducts with other aminoacids were detected.Therefore lysine residues within liver

membrane proteins are the major site oficetaldehyde adduct formation in vivo. Thevariable expression of the adduct maydetermine why only a proportion ofallcoholics develop cirrhosis.

A 139)3


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Activated complement is cytotoxic toisolated rat hepatocytes

A J K WILLIAMS, S K MOUI E, J I) MCGIlVAN, ANDR E BARRY (Department of Medicine, BristolRoyal Infirmary, Bristol) Activated comple-ment forms a terminal membrane complexwhich can lyse susceptible cells. Comple-ment activation has been shown to occur invivo in PBC, and we have shown thatacetaldehyde modified hepatocyte mem-branes will activate complement and thismay be hepatotoxic in alcoholic liverdisease. We have therefore studied theeffect of activated complement (AC) uponhepatocyte viability as assessed by cellularATP levels and gluconeogenesis. Isolatedrat hepatocytes prepared by the method ofKrebs (1974) were incubated in Krebs-Henseleit bicarbonate buffered medium,pH 7-4, 37°C±fresh sera with complementactivated via the classical pathway (by pre-incubating with anti-albumin antibody,37°C, 30 min). ATP levels were significantlyreduced in the hepatocytes exposed to ACcompared with controls (not exposed tosera) after 20 mins incubation (from11-8±0-79 nmol ATP/mg hepatocyte pro-tein, mean±SEM, n=5 to 7-5±09-, n=5,p<0'0I) and after 4() mins (6-8±0-77, n=5).Sera without prior activation of complementcaused no significant reduction in ATP(I 1-5±10-(, n=5 at 20 mins, 11-8+±0-79 at 40mins). Gluconeogenesis was reduced in thehepatocytes exposed to AC (2'08±t)05 nmolglucose/mg protein/min, n=4) comparedwith controls (4-7±0-33, n=4, p<0(0l).Activated complement is cytotoxic toisolated hepatocytes and could therefore bean important cause of hepatotoxicity invivo.

Is sympathetic activity really increased incirrhosis with ascites?

A J MACGILCHRIST, L G HiOWES, C HAWKSBY, J

REID, AND T J 1IIOMSON (University Depart-ment of Materia Medica, Stobhill Hospital,Glasgow) The increased plasma nore-adrenaline (NA) recently reported in severecirrhosis may be due to increased spilloverinto plasma from sympathetic nerves (andby implication, increased sympatheticactivity), or to reduced clearance of NA dueto reduced hepatic metabolism. We havedeveloped a new method to calculate NAspillover and clearance by intravenousinfusions of sub-pressor doses of tritiatedNA. We studied 14 cirrhotics with ascitesand 13 age-matched patient controls. Allsubjects studied were in-patients, off

diuretic therapy and receiving 40 mmol ofsodium/day. The cirrhosis was alcoholinduced in all but two cases, but no subjectconsumed alcohol within one wcek of thisstudy. Values wcre median (rangc) andcomparison was by Mann Whitney U test.Thus the clearance in plasma NA in

cirrhosis+ascites represents a markedincrease in global sympathetic activityrather than any reduction in clearance. Thiscould partly explain both sodium retentionand renal vasoconstriction, and favours the'underfilling' theory of ascites formation.

Oxygen free radicals do not activate thezymogens of human pancreatic proteases

P M GUYAN, J BUll ER, AND J M BRAGAN/A (TlieMedical School, University of Manchester,Oxford Road, Manchester) The currentinterest in superoxide (04 and hydroxyl(OH) free radicals as the initiating mecha-nism in pancreatitis, led us to investigatetheir effect on pancreatic proteases,zymogens and trypsin inhibitor. Equal con-centrations of these free radical specieswere generated, in different doses, using acaesium source (0.5-13 kraids) and a linearaccelerator (3-15 krads). Using specificsubstrates and sensitive fluorometric assaysit was confirmed that trypsinogen, chymo-trypsinogen and proelastase in human pan-creatic juice were not activated uponirradiation, whilst trypsin inhibitor was alsounaffected. A dose dependent (() 5-2krads) loss of trypsin activity was observedin dilute aqueous solutions of the pureenzyme, but not in activated human pan-creatic juice (doses 0(5-15 krads). whereasin the juice there was a dose dependent lossof chymotrypsin activity.These in vitro studies indicate that if O,.

and/or OH are involved at the initiationstage of pancreatitis, they exert theirdamage by mechanisms that do not involveactivation of pancreatic protease pre-cursors. This resistance to oxidative stresscould be due to the presence, in pancreaticjuice, of a recently described novel enzymethat scavenges hydrogen peroxide.

Carbohydrate sequencing of a new pan-creatic cancer glycoprotein marker

C K CIIING AND JONATIIAN M RHODIS(University Dept of Medicine and WaltonHospital, Liverpool) We have recentlyidentified a 3-5x 10)6 D serum glycoproteinthat is highly specific for pancreatic canccrand distinct from previously described

caincer markers (Gut X1987: 28: A372). Wehave now sequenced its carbohydrate sidechains by sequential degradation and lectinbinding.The undegraided piancreaitic cancer glyco-

protein binds the lectins peanut agglutinin(PNA, gal 1-3 gal NAc binding), ulexeuropacus I (UEA 1, fucose), limax flavus(LFA, sialic acid) and whecat germ agglu-tinin (WGA, sialic acid, NAc glucosamine).Serum containing the marker was electro-phoresed (SDS-PAGE. 2-1604,) and blot-ted onto nitrocellulose paper. The blot wassubjected to mild acid hydrolysis (5O mMH,SO4) followed by Smith degradation(sequential sodium periodaite (75 mM),sodium borohydride (0( 1 M) and sulphuricacid (25 mM)). Peroxidase tagged lectinswere used to analyse terminal carbohy-drates aftcr eaich step. Mild acid hydrolysisconsistently eliminates LFA and UEA Ibinding. Smith degradltion then causes lossof PNA binding with retention of WGA+.The pancreatic cancer marker epitope

therefore comprises:

galactose Ii 1-3 NAc galaictosiamine-R.

fucose sialic acid

Persistence of WGA+ on epitope negativechains after Smith degradation implies thatexpression of the marker results fromincomplete glycosylation.

Difluoromethyl-ornithine (DFMO) inhibitsadaptive ileal mucosal hyperplasia afterpancreaticobiliary diversion (PBD) in the rat

T BAMBA, S VAJA, G M MURPHY, AND R Ii[)OWL IN(. (Gastroeniterologyv Uniit, Guy'sCampus, UMDS of Guy's atid St Thomas'Hospitals, Lonidoni) Pancreaticobiliarydiversion, achieved by transposing 5() cm ofjejunum to lie between pylorus andampulla, stimulates ileal mucosal hyper-plasia but the role of polyamines and relatedenzymes (ornithine decarboxylase: ODCand diamine oxidase: DAO) in this adaptivereponse, is uncertain. Since ODC is inhib-ited by DFMO, we studied the effects of2%DFMO in drinking water on indices ofmucosal mass, ODC, DAO, and alk phosactivities and polyamine levels in six groupsof rats studied two weeks after surgery: (i)transected controls (TRC), (ii) TRC+DFMO (starting three days pre-op), (iii)TRC pair-fed with group (ii), (iv) PBD, (v)PBD+DFMO and (vi) PBD pair-fed.

Ornithine decarboxylase activity (nmolh 10cm ')increased from 2'60+SEM 0-47(TRC) TO 6X87±2-38 (PBD; p<0)-()5) but

A 1394


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


DFMO markedly inhibited ODC to0-45±00-5 (TRC; p<0(-(0) and 0(75+±021(PBD; p<0(002): pair-feeding had no sigeffect. There were corresponding changesin putrescine and spermidine but not inspermine. Alk phos activity increased 4-6fold after PBD but was unaffected byDFMO or pair-feeding. Conversely, DAOactivity fell after PBD in both ad-lib andpair-fed groups but did not change withDFMO. Difluoromethyl-ornithine inhib-ited PBD-induced increases in mucosal wetwt, prot and DNA/10cm by34%, 31%, and47% respectively (p<0(05-0(0 1).The results confirm that ODC and poly-

amines play a vital role in the cellularcontrol of intestinal adaptive growth.

Is continuous ambulatory monitoring ofduodenogastric reflux feasible? In vitroassessment of an enzymatic amperometricbile acid electrode

D ARMSTRONG, B LENNOX, J ALBERY, G M

MURPHY, AND R H DOWLING (Gastroenter-ology Unit, Guy's Campus, UMDS andDept Physical Chemistry, Imperial College,London) Duodenogastric reflux (DGR) hasbeen implicated in the pathogenesis ofgastric mucosal damage. Current DGRquantitation techniques provide only inter-mittent or short term monitoring of anepisodic phenomenon. Our aim is con-tinuous 24 h ambulatory monitoring ofDGR, using intragastric total BA concen-trations ([BA]) as a marker of DGR -

analogous to pH monitoring in the investi-gation of reflux oesophagitis. Any [BA]monitoring technique must be sensitive,specific, stable and able to detect all intra-gastric BA. We therefore used the estab-lished enzyme reaction as the basis for a BAelectrode. A platinum disc, coated with aconducting organic salt (NMP+-TCNQ ), iscovered by a membrane enclosing the 3-caOH steroid dehydrogenase. 3-ct BAs areoxidised by the enzyme, producing NADHwhich is then oxidised by the NMP' -

TCNQ , returning NAD' as substrate forthe enzyme. Oxidation of NADH generatesa current proportional to [NADH], andthus to total [BA] in solution. The semi-permeable dialysis membrane controls BAentry and prevents enzyme loss. We des-cribe the in vitro assessment of this elec-trode using solutions of chenodeoxycholicacid, deoxycholic acid and their glycine andtaurine conjugates (0-1-2-0 mM). Maxi-mum plateau current was generated witheach [BA] within three minutes and

remained stable for at least 10 minutes. Theelectrode showed a curvilinear current v

[BA] relationship (at I mM, 0(61±SEM0-04 ptA/mM, n=6). Over 30 hours therewas a fall off in electrode response (at1 mM, 0(34±00-3 [tA/mM) attributable todecreased enzyme activity. The electrodewas then compared with the standardspectrophotometric assay in determiningthe total [BA] in methanolic extracts of 10gastric aspirates, good agreement beingobtained (r=-)99, p<0-001). With furtherdevelopment, including miniaturisation andcomputerised correction for enzyme decay,this electrode should prove suitable for invivo monitoring of intragastric BA activityand DGR.

An integrated study of gall bladder (BG)motor function in gall stone patients

P PORIINCASA, P MITRA, M MA(ihISOULD[OO, G M

MURPIIY, AND R 11 DOWILING (Depts ofMedicine and Pathology, Gucy's Camnpus,UMDS of Guy's and St Thomas Hospitals,London) Gall bladder emptying in responseto exogenous or meal stimulated CCK isimpaired in gall stone patients but fewstudies have related cholecystokinetics toquantitative histology of muscle mass or thedegree of inflammation in the GB wall.Therefore in longitudinal strips of freshGB's, we studied tensiometric contractileresponses to CCK-OP (0-875x 10"' to 10M) and acetyl choline (ACH: 5x 10 'to 10(M), recording the max tension produced,and relating the dose required to generate50% max response (Ds,) to muscle area(MA'/O of total cross sectional area) and aninflammatory score (6 parameters).

In 14 patients with moderate chroniccholecystitis there was myohypertrophy,the MA% of 15-0±SEM 1-43 being greater(p<0.001) than that in four histologicallynormal GB's (7-72+0-24). The MA%0correlated (r=0-62; p<0-005) with the CCKDi,, which was also greater inWinflamed(3-45±0+006) than in normal (0-54±0 1t)nM) GB's but less ACH was required forthe DR (6-38+1-3 v 15-0±9-9 FM) -

perhaps because of denervation hyper-sensitivity. The max tension (kg/cm2muscle) in response to CCK was compar-able in normal (0-29±0.05) and inflamed(0-31+0-05) GB's but after ACH, it Wasgreater in normal (0-35±0-15) than incholecystitis (0- 19±0-02).These results confirm our in vivo findings

and suggest that despite myohypertrophy,the inflamed GB is less responsive to CCK,but more sensitive to ACH, than normal.

Restitution of colonic mucosa in vitro

P H ROWE., N 1 K FA(i(G, ANI) R C MASON

(Departments olf SurgerY and{ Pathology.,UMDS, Gi s's Hospital, Londtloni) It haspreviously been reported that completerestitution of epithelial integrity inamphibian gastric mucosai occurs in vitrowithin four to six hours after injury by IMNaCI. The process of restitution has notbeen previously described or investigaited inthe colon. Using chambered bulIlrogcolonic mucosae (n= 10) exposed to IMNaCI in the luminal chamber exhibited ainimmediate fall in potential diff'erence (PD)from baseline values of' -58-6±6-3 mV(±SEM) to 0 mV (p<0)-05) and a fall inresistaince (R) from 216± 18-3 Q cm' to8±4-8 Q cm' (p<0-05). Tissues removed(n=-5) following NaCI exposure exhibitedsevere surface mucosal injury with denudai-tion of the basal lamina. After 10 minuteNaCI exposure, and replacement of NaCIwith Ringer's solution, at luminal pH 5()when tissues (n=5) were allowed to recoverfor four hours in the chambers, there washistological reconstitution of the surfatceepithelium and the PD and R had returnedto -43±4'5 mV and 220±14 2 Q cm'respectively, not significantly different fromthe PD and R observed in control tissuesmaintained for four hours in the chambers.Control tissues (n=4) showed no significantchange in PD or R and no histologicalevidence of damage.We conclude restitution of the amphibian

colonic mucosa occurs in vitro followinginjury by IM NaCI.

Effects of starvation and refeeding onelectrogenic ion transport in the rat colon: amodel for famine diarrhoea?

11EL [N NZEGWU, A YOUNG, AND R J l.-VIN(Department of Physiology, University ofSheffield, Western Rank, Sheffield) Theaetiology of the diarrhoea induced byfamine and by subsequent refeeding ofvictims is unknown. In the rat model, star-vation hypersensitises the small intestine tosecretagogues resulting in greater fluid andelectrolyte secretion, but little is knownconcerning the colon. Male rats were pro-gressively starved for up to 72 h and thenrefed for up to a further 72 h. Electrogenicion transport was measured as the shortcircuit current (Isc) under basal conditionsand with bethanecol (I mM) stimulation.

Basal Isc fell during starvation, a maxi-mum decrease of 18% (p<0-05) occurringafter 72 h (n-=29) compared to fed controls

A13(95


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroemterology

(n=26). It then increased over the 72 hrefeeding period until it exceeded the fedlevels by 18% (p<0-05, n= 12). Bethanecolstimulated A Isc (A Isc=Max Isc- BasalIsc) increased significantly after 48 h ofstarvation (+40%, p<0(0l, n=10) andremained at this raised level. On refeedingit returned to the fed level only after 72 h(+ 12%, n= 12, p>0'05). These results maypartially explain the famine diarrhoea andthe diarrhoea associated with realimenta-tion in man. Support from the BritishDigestive Foundation is acknowledged.

Atrial natriuretic peptide influences electro-lyte transport in the large intestine

K J MORIARTY, N B HIGGS, M LEES, A IONGi, G

WARHURST, AND 1 A IURNBERG (DepartmentoJ Medicine, Hope Hospital (University ofManchester School oj Medicine), Salford)Atrial natriuretic peptide (ANP) possessespotent diuretic and natriuretic propertiesand appears to play a central role in fluidand electrolyte homeostasis by an action onthe kidney. We examined the influence ofANP on the large intestine, because thisorgan is also intimately involved in homeo-stasis. Segments of distal colon of maleSprague-Dawley rats were stripped ofmuscle layers and mounted in flux cham-bers. Atrial natriuretic peptide, whenadded to the serosal aspect of the mucosa, inconcentrations ranging from 10 t-1() ' M,caused a rapid, dose-dependent rise inshort-circuit current (Ij,) and transmucosalelectrical potential difference (PD). Atrialnatriuretic peptide (11) M) elicited a peakincrement in I, of 59-2±6-7 ,uA/cm' (n=8,p<0()(l) and in PD of 2-97+±)3 mV (n=8,p<0(00l). A rise in I,. and PD is generallyassociated with enhanced secretion. Theresponse to ANP (It) M) was, however,virtually abolished by pre-treatment withthe calcium channel blocking drug, d,l-verapamil (It) M mucosally and serosally),and also by using a calcium-free (nocalcium +2x 1)0 ' M EGTA) solution on theserosal surface.These findings support a role for ANP in

the regulation of colonic ion transport andsuggest that ANP may elicit secretion via acalcium mediated mechanism.

SMALL INTESTINE

Somatostatin analogue SMS 201-995 in thedumping syndrome: effect on packed cellvolume (PCV)? Blood glucose (BG) andplasma GI peptides

J N PRIMROSEI, D Jo(HNSI)ON, C SHAW, AND) K I)

BUCHANAN (University Departtnent ofSurgery, Leeds General InfirmnarY antdDepartment Metabolic Medicine, Queen.sUniversity, BelJast) The aim of this studywas to determine the effect of SMS on thedumping syndrome (DS) provoked by

hypertonic glucose, and on the accompany-

ing changes in PCV, BG, and GI peptides(insulin, N-terminal-glycogen-like immuno-

reactivity (N-GLI), GIP, VIP, neurotensinand N terminal neurotensin immunoreact-ivity (N-neuro)).Dumping was provoked in 10) patients

with DS by means ot 350) ml 25%/ glucose.Patients received pliacebo, SMS 50) ug or

SMS 100 ug, given in random order on

separate days. Blood samples were takenfor PCV, BG, and GI peptide estimationsthroughout the two hours of the test.

With placebo, all patients experiencedsevere symptoms: nine had early dumpingand thrce, late dumping. Symptoms were

abolished or much reduced aftcr either doseof SMS.

Somatostatin analogue is effective in

treating the symptoms and reversing thechainges in PCV and BG of' dumping pro-

voked by hypertonic glucose. The incre-ments in GI peptides are reduced signific-antly by SMS, but only the change in N-GLIcorrelates with the change in PCV (RS=0(82, p<0'05). N-GLI may reflect, or be

associated with, the changes of dumping.

Prostaglandin (PG) E, is a mediator of 5-hydroxytryptamine (5-HT) induced fluidsecretion in the human jejunum

1. K MUNCK, A MlERT/-NIIF SIS.N, Ii WlSI I, K

BUKIIAVIE, 1. BI UBILIER, ANI) J RASK-MAI)SIFN

(Depts of Medic al Castiroenlerologv,Bispebjer,g, atnd Herlev Hospials, Univ ofCopenhalgen, Dentmlark anzd Ep C('liiiPharmaccol Univ ( CGraz, A lstriai) Studiesin the rat jejunum in vivO have shown that5-HT causes fluid secretion accompanied byluminal PGE, releatse. These effects cain heblocked by indoomcthacin aind kctalnscrin,suggesting that PGE, may he ain importantintermediate in the transduction mecha-nism leading to 5-HT induccd secretion. Totest this hypothesis in man 'stcady state'perfusions (9 ml/tnin) of 31) cm proximaljejunum were performed in eight healthyvolunteers, using a triple-lumilen tube and a

Ringer's solution with ''Cr-EDTA as a

non-absorbable mnarker. 'Ihe eftects of

exogenous 5-HT (10) ftg/kg/min iv) on fluidtransport and jejunal flow rate (JFR) of

PGE., before and after the aidministraition

of indomethacin (1-0 mg/ko iv), weremacasured in 15-min periods for 2x60 nimifollowing a 60) imilt conltr-ol period. 5-HIrcvcrsed fluid absorption (+ 3.9()-7 Inl/h/cm; mrcan±SE) into protusC sccrction(-1 7+±()5; p<0(0(l Duncan's multipicriange test) and significantly increaised JFRof PGE. (0 11 +01)(3 v () 40+() 1) ng/min;p<0(-0 ). Indomethacin piartly restoredfluid aibsorption ( + 1 2+05 p<0-(0) andinhibited JFR of PGE. (0)21 ±10))4p<0()05). Thcse rcsults providc furthcrevidenicc in support of the thcory that PGsare involved in intestinal fluid sccrctioninduced by 5-HT.

Human upper gastrointestinal transit andcarbohydrate absorption are modulated bybeta-adrenoreceptors

A S MCINIYRIE, 1) (i III)OMPS()N, W R BURNHIAM,ANI) F1 WAI.KE.R (Dept Gastroenterology, 7'TheLontdoti Hospital (VWhitechapel), LondonantI Oldchurch Hospital, Rotn/ord.FrdEssex)(Control studies: O)rocaecal transit ol'a souplaictulose metal was measured in 26 subjectsby the serial hreaith hydrogen method.Ciarbohydrate athsorption was estimnatedwith at series of fructose meals, fructosebeing varied in 10) g increments (2(0-(90 g);the smallest dose producing ai hydrogen riseindicated that the absorptive theshold wiasjust exceeded. Transit wals consistentwithin, but varied widely between indi-viduals (median 55. range 32-172 min).Fructose aibsorptive threshold showedsimilar interindividuali variation (mediain40) range 30)-8ff g and correlated withtrIansit (T-().70, p<0-0() 1

B-adrenergic stimulation: Isoprenaline(Bl alnd B2) consistently inhibited transit(mean A 30-4+3-7 mnins, p<0(-00l) aindincreased fructose absorption (mean A 15g), the magnitudc ol response varyinginversely with control data (T=-) 52.p<0(0 I ). Atenolol (B I antagoniist)abolished the cttfct.

B-adrenergic blockade: B-blockade (Band B2 or B I ailone) aiccelerated transit(mean A 21 7+7 4 mins, p<(0)2) aindreduced tructose absorption (meiian A 15 g);the effect wias directly reliated to controltrainsit (T=0()49, p<0(02).Thus a tonically active B -Iadrenorecep-

tor palthway moduliates normal nutrienttriansit iIIad carbohydrate aibsorption alndiaiccounits for up to 67(YO of differenccbetween individuials.

Characterisation of defective sodiumcoupled glucose transport in congenital

A 1 396


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


glucose-galactose malabsorption (GGM),using jejunal brush border membranevesicles

I W BOOTI1, D SULE, P B PATEI, G A BROWN,AND K BEYREISS (Institute of Child Health,University of Birmingham and Karl MarxUniversity of Leipzig, Germany) To date,the defect in D-glucose transport seen inGGM has not been directly shown in thejejunal brush border membrane. We havetherefore prepared brush border membranevesicles from jejunal biopsies from a threeyear old with GGM, and from nine controlchildren, using a miniaturised differentialcentrifugation technique. Diagnosis in thepatient was confirmed by jejunal steady-state perfusion studies, which showeddefective glucose and galactose absorption(p<0-((l), but normal fructose transport.In brush border membrane vesicles fromcontrols, an inward Na+ gradient (100mmol/l) enhanced 15 sec D-glucose uptakex6-5 (p<O-Ol). Those from the patientshowed not only a 900/o reduction in Na+-coupled D-glucose uptake at 15 sec(p<0-001), but also an elevated passiveuptake (p<0(005) with equimolar Na' inboth intra- and extra-vesicular spaces. Na'/H+ exchange was normal.These studies confirm for the first time

the presence of grossly defective Na -coupled D-glucose transport in the jejunalbrush border membrane in GGM, andsuggest that brush border membrane vesiclestudies may be the optimal technique forconfirmation of the diagnosis.

Phosphatidylcholine (PC) degradation insmall intestinal bacterial overgrowth

M HEALY, M P O CONNOR, C T KEANE, D G WEIR,AND R R O'MOORE (Depts of ClinicalMedicine, Clinical Microbiology and Bio-chemistry, St James' Hospital and TrinityCollege Dublin, Ireland) In vitro experi-ments with anaerobic bacterial strains, fre-quently found in small intestinal bacterialovergrowth (SIBO), show enzymaticactivity capable of degrading the cholinemoiety of PC to the aliphatic amines (AA).trimethylamine (TMA) and dimethylamine(DMA). Since PC plays an important role inlipid absorption, a study was undertaken toassess the PC degradation parthway, using arat self-filling blind loop model (SFBL,n=10 and controls, n=6) and SIBOpatients.Animal study: Urinary excretion of AA's

was measured on a choline free diet (PC 2g/kg dry food). SFBL DMA excretion was

greater than controls (14±2 v 5±0-9 [imol/day, x± SE, p<0-005). Antibiotictreatment reduced SFBL DMA from13-1±2-2 to 3-8±0+6 [tmol/day. p<0(02. Toconfirm the pathway from PC to AA,orogastric intubation was performed with(N-methyl-'H) PC and -H-DMA identifiedin the urine.Humani study: 10 controls and 12 SIBO

patients (jejunal culture>l05' orgs/ml) weregiven a fatty meal to stimulate biliary PC.Urine 4-hr AA excretion was measured andresults expressed as total AA (DMA+TMA)/creatinine ratios. The excretion ratiowas greater in SIBO patients, 98±22 v 33±3(controls), p<0-02. This study indicatesgreater PC degradation in both the animalmodel and SIBO patients. Antibiotic treat-ment reversed this effect in the animalmodel. The bacterial degradation of PCmay be a factor in the pathogenesis of thesteatorrhea in SIBO.

Influence of orally administered amino acidsand peptides on protein turnover kinetics inthe short bowel syndrome

R G REFS, G K GRIMBI , D HA!I DAY, C FORD,AND D B A Sil.K (Department of Castro-enterology - Central Middlesex HospitalLondon, Acton Lane, London and ClinicalRe.seurch Centre, Harrow) Experimentalevidence supports the use of peptides ratherthan amino acids (AA) as the N source inelemental diets. In this prospective random-ised cross-over study, two isonitrogenous,isocaloric diets with identical AA composi-tion were administered to four mal-nourished patients with the nutritionallyinadequate short bowel syndrome (<100cm). Each received a five day 'run in' periodof enteral nutrition and were then random-ised to receive a diet containing either apartial hydrolysate of casein (PEP) or theequivalent mixture of free AA. After afurther five days the alternative diet wasgiven and on completion of each studyperiod, whole protein turnover wasmeasured by a primed-dose, constantinfusion of 13 C-leucine. There was nosignificant difference in the rates of flux,synthesis, breakdown or oxidation duringadministration of the PEP as compared withthe AA diet. Cu (II) - Sephadex chromato-graphy showed that 50% of alpha amino Nin the PEP diet was present as tetra- orhigher peptides. These findings indicatetherefore that in the short bowel syndromethere are no apparent short-term nutritionaladvantages of administering a diet whosenitrogen source is composed of mediumchain length peptides.

Application of organ culture to the investiga-tion of lectin-induced damage to the brushborder membrane

R M BA1, C'AHART. 1 MC(IiAN,13 GFiT.l ANDJ R SAUN DF RS (Der)(trllnent.s of VeterinarYPatholog,', Medical Microbiology (1a1ndMicrobiolog,y, Universitv of Liverpool)Ingestion of lectins such as phytohaiemalg-glutinin in red kidney beains is an importaintcause of food poisoning, but the patho-physiologicill mechanisms ire poorly under-stood. Lectins bind to specific carbohy-drates in brush border glycoproteins, butthe relationship between binding and lectin-toxicity is unclear. In this study, the ultria-structural effects of lectins on the brushborder have been exaimined during orgainculture and related to their receptorspecificities.

Explants from ratbbit ileum were culturedfor 24 hours in medium containing cach of10 lectins, and were then processed forelectron microscopy. Phytohiaemaigglutininand lectins with specificity for N-atcetylglucosamine or N-acetyl neuriaminic aicidcaused distinct damage to the brush bordermembrane at concentrations betwecn 10and 100(tg/ml. The most constant featturcwas elongation and vesiculation of micro-villi which in some cases appeared to bejoined aIt the tips. In contrast, lectins withspecificities for glucose, mannose. galactoseor N-acetyl galactosamine had no discern-able effects even at 200()g/ml, a concentra-tion considerably higher than that requiredfor haemagglutination.These findings provide evidence for ultra-

structural damage to the brush border mem-brane by specific lectins. The apparentrelationship between damage and receptorspecificity may reflect either accessibility ofappropriate binding sites or a differentialresponse to binding.

Postoperative pain relief with diclofenacsuppositories

R F It DAWE-S AND G i ROYILE (UniversityDepartment of Surgery, SouthamptonGeneral Hospital, Southamptoni) Diclo-fenac is a non-steroidal analgesic and has along half life (12 hours) when given as asuppository. The purpose of this study wasto investigate its role as a post-operativeanalgesic for abdominal surgery. Sixty ninepatients having elective abdominal surgerywere entered into a prospective randomiseddouble blind trial of diclofenac (Voltarol,Geigy) as an adjunct to standard im opiatepostoperative analgesia. Patients received

A 1 397


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society ofGastroemiterology

eight suppositories over four days- that is,12 hourly, containing either 100 mg diclo-fenac or inert substance. Abdominal painwas recorded by the patients regularly on a10 cm visual analogue scale (VAS) afteroperation, both 'at rest' and after coughing.Patients were closely monitored for anyunwanted symptoms or complications, andall drugs administered were recorded.Thirty one diclofenac and 26 placebopatients fully completed the study. Dailypain scores measured 'at rest' were similarbetween groups. On coughing diclofenacpatients experienced consistently andsignificantly (p<0c02) less pain on all thepost op days. Opiate analgesic consumptionwas similarly reduced in the diclofenacgroup. There were no differences in post-operative complications between the group.It is concluded that diclofenac is of use inpost abdominal surgery pain relief whengiven 12 hourly as a 100 mg suppository.

SMAILL INTESTINE

Use of endoscopic biopsies to diagnosedisaccharidase deficiencies

J A SMITH, J AMOAH, D O REILLY, J F MAYBERRY,AND R G LONG (Medical Research Centre,Citv Hospital, Nottingham) Endoscopicduodenal biopsies can replace Crosbycapsule jejunal biopsies for the histologicaldiagnosis of malabsorption despite thepresence of shorter villi and Brunner'sglands. As endoscopic biopsy is quicker andeasier to perform, we compared disacchari-dase levels in endoscopic biopsies from thesecond part of the duodenum and Crosbycapsule biopsies from the proximaljejunum. Thirty patients had normal histo-logy and 12 patients had partial or totalvillous atrophy. Two normal patients hadprimary hypolactasia diagnosed by bothbiopsy techniques. Mean+SEM lactaselevels in 28 normal duodenal biopsies were12 1±2-3 U/g protein, and in jejunal biop-sies were 21-2±4-0 U/g protein (p<0-()1). Aparallel statistically significant reduction inthe duodenum was seen for maltase andsucrase results. In villous atrophy secondarydisaccharidase deficiencies were found andcorresponding lactase results were 200+06and 4-0+1-2 U/g protein. Thus normaljejunal levels are consistently higher thanduodenal levels but results in villousatrophy are not statistically different. Weconclude that duodenal biopsies are a validalternative to Crosby capsule jejunal biop-sies for the diagnosis of both primary andsecondary disaccharidase deficiencies.

Concomitant presence of IgE plasma cellsin duodenal and bronchial mucosae inasthmatic patients

C ANDRf, [) GINDRE, Y PACHIECO,.I)1DSCOS,ANt) M PERRIN-FAYOLILE (Laboratoired'Inmunopathologie Digestive INSERM,Ceentre Hospitalier Lvoon-Suld, Pierre-Benite, France) Immunoglobulin-containing cells of all classes were countedaccording to the tissue unit method induodenal and bronchial mucosae in controlsubjects and asthmatic patients.

In control subjects (n=6) the mean per-centages of IgA, IgM, IgG, IgD, and IgEcells were 74, 8, 17, 1, and (O respectively inthe bronchial mucosa. In 25 control subjectsthe mean percentages of IgA, IgM, IgG,IgD, and IgE cells were 75, 13, 10, 1, and 1respectively in the duodenal mucosa. Thesame studies were performed in 15asthmatic patients. The differential countsof IgA, IgM, IgG, and IgD cells did notdiffer from the values observed in controls.A significant increase of IgE cells was foundin the duodenal and bronchial mucosacfrom nine patients, the average values being13'% in the digestive mucosace and 24%S inthe respiratory mucosa. No IgE plasma cellswere observed in the mucosae from fiveother patients. In the last patient there wereno IgE cells in the intestinal mucosa butthese cells accounted for 18% in therespiratory mucosa.The similarities of the local allergic

response in asthmatic patients may be inter-preted in three ways. Food allergy may beinvolved in many asthmatic patients. Theimmune response could be the consequenceof an allergic stimulation at both levels. Theobserved data also correspond to the con-cept of an universal mucosal immunesystem: cells stimulated at one mucosal sitemay migrate to other mucosal sites.

Feedback regulation of transit and motilityin the human jejunum and ileum by ilealinfusion of glycochenodeoxycholic acid(GCDC)

R PENAGINI, R C SPILLER, J J MISIEWICZ, AND

P G FROS'I (Depts of Gastroenterology anidNutrition, and of Chemical Pathology,Central Middlesex Hospital, London) Thepresent study examined the possibility thatpresence of GCDC in the human ileum inamounts comparable to a bile acid pool (5mmol) inhibits fed ileal and jejunal motility.

T'en healthy subjects were studied.Transit times (by a marker bolus of bromo-sulphthalein), intraluminal pressure

activity, flow rates (by "4C-PEG) and intra-luminal bile acid concentrations (BA) weremeasured in a 40 cm jejunal segmentproximal (n=6) and a 40 cm ileal segmentdistal (n=6) to an ileal port infusing anisotonic clectrolyte solution with or withoutGCDC for 80 min. During GCDC infusion(6(0 [tmol/min) jejunal and ileal transit weremarkedly (p<0-05) delayed (31.6+7.7,mean±SEM. v 14 5±3-8 min and 37-)0±5.7v 21-0+3-5 min, respectively), while flowrates did not change significantly (4-6±0-6 v4 5+±06 ml/min and 4X8±0 5 v 4 2±0-3ml/min, respectively). Per cent durationpressure activity in the ileum decreased(p<0.05) promptly (60)±8 sec) after thestart of GCDC infusion (6.9% + 1-6 v17 9'Y.±4-2). Inhibition of jejunal motilitywas more gradual and reached significance(p<0(05) only 30 min after start of GCDCinfusion (18X9(X.±3-2 lo 35-2%±7-1). Intra-jejunal BA were not altered by GCDCinfusion, while intraileal BA increased(p<0.05) during GCDC infusion andshowed a 35-92% (median 45%) absorptionof GCDC by the 40 cm ileal segment.

In conclusion these observations suggestthe existence of a regulatory mechanism inhealthy man, whereby presence of GCDCin the ileum inhibits motility and delaystransit throughout the smaill intestine.

Wheat protein, gut abnormalities andrheumatoid arthritis

CLIONA ) FARRIEILY, I) MAR'I'EN, I) MELCIIER, RSF11ERWOOD, A J GOILDSTHIN. FERNANDESANI) B MACDOUGALI. (Schlool of BiologicalScienices, University of Su.ssex anid Depts oJHistopathology, BiochemistrY, Rheluma-tology, anid Gastroeniterology, Royal SuissexCountY Hospital, Brighton, Suissex) There isincrecasing interest in the role of the GI tractin the pathogenesis of rheumatoid arthritis(RA). We studied this role in 87 RApatients (65 women; mean age 62). UsingELISA, 50 of the 87 had raised levels of IgGto wheat and/or milk protein (41 werepositive for wheat, six for milk and 12 forboth). Ninety per cent of the antibodypositive group had raised levels of IgArheumatoid factor (RF) compared with only27°/ (p<0-00l) of antibody negativepatients (both groups were on similar treat-ment regimens).

Jejunal biopsies were ciarried out on 25 ofthese patients: 15 had raised levels of bothIgA RF and wheat protein IgG (AB+); theremaining 10 had normal levels of bothantibodies (AB- ). The biopsies were

A1398


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


assessed using standard histological, stcreo-logical and biochemicail techniques. VillousaItrophy was present in six of the AB + but in

only one of the AB- patients. This wIsconfirmed by ai lower villous surface/volumcratio (77-8 17-3 SD) in AB+ patients whencompared with AB- (97-9±21 8: p<0)-02)or 10 age-matched control piatients (125+17.9; p<).-00 1). Lactase levels ( meaisured as

umol/min/g tissue) were ailso lower in theAB+ group (34±2 () SD) than in the AB-(5-7±2-5; p<t).-)3) or control groups

(6-4±2.7; p<)-00)l).These findings of GI abnormalities

associated with raised levels of IgA RF aindwheat protein IgG in our patients suggestthat the gut may he involved in the immuno-patthogenesis of RA.

Familial visceral myopathy

( A RODRIGUES, N A SiiiHPiIERt), I' R IlAWITi Y, J

1 I ENNARD-JON1-S, ANI) If H iIIOMPSON (St

Mark's Hospital, Cit'v Rotid, LoIndoo, TheLondotn Hospital (Whitechapel), London)The first British kindred with autosomaildominant viscerail myopathy is described:Six affected members in two generations all

had intestinal pseudo-obstructioni whichwas demonstrated by conitraist radiology.Five of these patients had ai megaiduo-denum, two haid diminished oesophagealperistalsis, four small bowel aind fourcolonic involvement. One patient haddilatation of one ureter and incompleteblaidder emptying on intravenousurography.The clinical onset of the disease occurred

in childhood or aidolescence with dysphagia(n=-2), abdominal pain (n=5). abdominaldistension (n=6). constipation (n 4),diarrhoea (n=4), aind urinary symptoms(n=2). The two most severely affectedindividuals arc on home parenterill nutri-tion and have undergone surgery for thealleviation of symptoms. Three otherpatients have mild, non-progressivesymtoms. and have been able to maintainnormal nutritionail status without special

measures. The sixth family member hasbeen lost to follow up, and his current statusis unknown.The main pathological features consisted

of marked dilatation alnd thinning of thebowel wall, and vacuolar degeneration ofthe longitudinal layer of the muscularispropria with fibrosis and elastosis. Thesubmucosal and myenteric nerve plexusesappeared normal and there was no histo-logical abnormality ot the blood vcssels.

Is long term home parenteral nutritionworthwhile?

M A STiKES, J 1 SiiAF1IFR. M 11 IRVIN(, (ONBEiiAI O')F i111. UKII I(MIi PARiNiALRAIN RI lI ION (ROUP) (Departmnent.s ofMedicinea(1d(1 Surgersy, University ol Ma,l-chester School oJ Medlicinie, Hope Hospital,SalJord) There are 241 patients entercd on

the HPN register for the United Kingdomand Irel.ind. Fifty two of these haive beenl onHPN for two or more yeiars. There aire 33women and 19 men with an avcrage agoe of36 years (range 4-60). Over half have been

entered on the register by the two largestcentres. In fact, five centres aiccount forover 80% of these long-tcrm HPN patients.

Crohn's disease is the underlyilng path-ology in over half, and 68% are on HPNbecause ot the short-bowel syndrome. Theyhave becn on HPN for a total of 19t) years.Forty one of themnare still on HPN, thelongest has been on tor seven years. Fourpaticents have resumcd enteral fceding, thelongest 41 months after commeclcing. Six ofthese patients have died, two after five yearson HPN.These longterm patients have a signific-

antly better lifestyle (p<()-()l ). with a lowerincidence of sepsis (0-25 per patient pcryear; p<0)-05), and a lower incidcnce oftotal complications (0(56 per patient peryear; p<0))(0l), than for the group of HPNpatients as a whole.Home parenteral nutrition is a highly

succcssful treatmcnt in the managemcnt ofpatients with longterm intestinal failure.

Jejunal mucosal architecture in HIVinfected male homosexuals

P A BAIMAN, A MtLL ER, S FORSTE1R, A PINCH-ING, W tIARRIS, AND (G F GRIFFIN (Departmentsof ('omnunicable Diseases tand Histo-p)athologs', St George's Hospital atnd StMars 's Hospital Medical Schools, LonIdon)Crosby capsule jejunal biopsies from HIVantibody positivc male homosexuals (n= 20:asymptomatic=five, persistent generalisedlymphadenopathy =nine, AIDS=6) weresubjected to light microscopy and stainedwith H and E, modified Zichl-Neelsen,Giemas, PAS and Gram (to detect mucosalenteropathogens). Mucosal surface area/volume ratio (S:V) was dctermined using

Wcibel graticule. Crypt length (CL) was

assessed by counting enterocytes from cryptbase to crypt/villus junction. Intraepitheliallymphocytes (IEL), detected usingimmunoperoxidase technique (leucocytecommon aintigen), were counted/500

centerocytes for cach biopSy. JeCju.nal hiop-sies f'rom aige matched maile control sublhjets(n=seven) were similalrly treated. Nomucosal eniteropathogens were detected inainly bliopsy. Enterocytes appeared normaland villous aitrophy (VA) with crypt hyper-plaisia wats the only abnormality detected atall staiges of FIIV disease; S:V (control)49-2+6-2, (HIV) 38-4+94, p=)-006; Cl(control) 32-5±6-8 (HIV) 38X8±75,p=0)-)498. Crypt lcngth also invcrsclycorrelatted with S:V, r=-0f7 p=-0))0005.Intraepithelial lymphocytcs (coiitrol)19f)0±65, (HIV) 24 5±7'9, p=NS(numerical results, menan± I SD).

'I'hus jejunal crypt hyperplastic VAoccurs ait any staigc oflHIV disease. Cryptlength correlates with the degree ot VA antdI El counts i re normal.

I'ANCRI.AS

Glycochenodeoxycholic (GCDC) andglycocholic (GC) acids infused in the humanjejunum inhibit small bowel motility andtransit

R PENAGINI, J J MISIiWICZ, ANI) G(i FROST

(D)epts of Gastroenterolog' anid N1utritiont,and of Chemical Pathology, Central Middle-sex Hospital, Lonldoni) The present studyexamined the effect of jejunatl infiusion ofGCDC and GC on small bowel transit time(SBIT: lactulose 5 g-hydrogen breath test),fasting jejunal motility aind serumtbile aicidconcentrattions in 18 healthy subjects.

Each subject was studied during jcjunalinfusion of saline ailone (control) aind duringinfusion of at least one of four saline solu-tions containing: (1) GCDC 3-3 mmol(n=5), (2) GCDC 3.3 mmol+lecithin 08-mmol (L) (n=5), (3) GCDC 5 rnmol+L(n=6), (4) GC 5 mmol+L (n=6). Eaichsolution was infused in X() min on a separateday in raindomised order. Fasting jejunalmotility was recorded in the experimentstesting (3) and (4) only. Blood samples werecollected at t), 30, 60, 9t), 120, 15t) min fromstart of infusion for measurcment of tottilbile aicids (cnzymatic method): integratedincremental response (BA IIR) wascalculated.

Results (mean+SEM) show that GCDC5 mmol+L and GC 5 mmol+L delayed(p<0-05) SBTT when compared with con-trol infusion (158X3± 12S5 min v 111-7+17-6min aind 103-3±21-8 min v 7()()+ 14.9 min),inhibited (p<0-05) the percentagc durationpressure activity of phasc 2 (131± 1-s'828X1 +3.4%Y0 and 2932±55.5(Y v 34-9±3-9%),

A 1 399t


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


but did not change duration of migratingmotor complex, or of its phases. Glyco-chenodeoxycholic 3-3 mmol+L had a lessconsistent effect on SB1T (p=NS). All fourinfusions containing bile aicids increased(p<0-05) BA lIRs in comparison withcontrol infusion.

In conclusion these observations suggesta role of endogenous bile acids in thereguliation of small bowel motility andtransit.

The histopathology and staging ofcarcinoma of the ampulla of Vater

I C IAILBOTI, J P NE'OPiOlFtMOS, I) (CARR-I(CKi.,ANI) I) SIIAW (Dept oj Patilologv anidSurgery, Clinical Science.s Blildling,Leicester RoYal Infirmntarvy, Leicester)Carcinomas of the armpulla are uncommonand data concerning their paithology andbehaviour aIre inconsistent aind of question-able reliability. We have analysed thetumour type, grade, paittcrn of local inva-sion and outcome in 26 carcinomas surgic-ally resected over a 13 year period. Twenty-five (96%/O) wete intestinal type adcno-carcinomais, the degree of differentiiationbeing good (eight), moderaite (13), or poor(four). Adenomca cocxisted with aideno-carcinoma in 1 1 cases, six ol these being welldifferentiated (low gradc). There wasepithelial dysplalsia of duct epitheliumadjacent to 10 carcinomais, only two of thesebeing low grade.

Forty five per cent of paitients survivedlive years. The fivc yeair survival raite for thewhole series (estimated by the Kaplan-Meier mcthod) was 520o and, with lowgrade adenocarcinomai, 83% (p<t)()03).Using a staging system, based on extent oflocal sprea1d (Stage I=invasion of commonbile duct wall only, II rinfiltration intoduodenal wall and/or retroperitonealadipose tissue, 1111= infiltration of pancrcas,IV=Iymph node metastases), longtermsurvival correlates inversely with stage,both ailong (p= )0055) and by multivariatcanalysis, independently of grade. A simplescoring system, combining both gradc aindstagc precisely predicts survival (p<00() 1).

Endoscopic management of inoperablecholangiocarcinoma using Iridium-192

R J EDi, A HiAiTFIiELD, S MCINiYRE, ANI) (G MAIR

(Departments of' Gastroenterology andtRadiotherapy, The London Hospital,London) Improved survival has recentlybeen reported in patients with inoperable

cholangiocarcinoma after intubation by thepercutaneous route and intralluminal radio-therapy. We report a less invasive endo-scopic technique in which the Iridium-192wire source is inserted down a nasobiliarycatheter placed within a previously insertedendoscopic biliary prosthesis. Six men,mean age 64 years (range 41 -80) withinoperable cholangiocarcinoma hiave so farbeen treated. Following biliary decompres-sion when the serum bilirubin had fallensubstantially radiotherapy was commenced.The Iridium wire was loaded into a 0))038"Teflon-coated movabie core guide wire andinserted through the nasobiliary catheterunder fluoroscopic control. A total dose of60(X) rads was aidministered over a meian offour days (range 3 3-48). During thisperiod serum bilirubin increased slightly inthree patients but fell again aifter removal ofthe Iridium wire. Mean hospital stay afterthis treatment was 6-3 days (range 1-26).The only complication was mild cholangitisin two patients. Two patients have sincerequired stent replacemcnt. One patienthas died 105 days after treatment: theremaining five are all alive bctween 10( and430 days. This technique was well toleratedby all patients and in view of the reducedmorbidity of endoscopic compared withtranshepatic prosthesis insertion should beconsidered in patients with inoperablecholangiocarcinoma.

Morphological changes in the pancreas,suggestive of endocrine regeneration,induced by transplantation of isolatedhepatocytes

V JAFE'I, II DARBY, AND) H J FlIIODGSON (Deptsof Surgery and Medicine, Roval Post-graduate Medical Sclhool, HamnmersmithHospital, London) We have shown thattransplanted isolated syngeneic rat hepato-cytes grow within the pancreas. Liver cellsare found in interlobular spaces and pern-ductular sites up to three months afterimplantation, and after this a remarkableassociation is established between liver cellsand the endocrine pancreas - hepato-cyte-islet rosettes - a halo of liver cellsaround a core of islet tissue.

This report highlights the atypical pan-creatic morphology in these animals whichappears related to the presence of hepato-cytes. '[he principle changes include anincrease in number and size of islets, andregions of diffuse neuroendocrine cells inassociation with areas of ductular radicles.Histochemical staining has shown theseextensive atypical areas to contain function-

ing endocrine tissuc. The appearances aresuggestive of pancreatic endocrine neo-genesis and regeneration. When combinedwith a surgical procedure to divert portalblood across the pancreas longterm, thepancreatic changes associated withimplanted hepatocytes are even moremarked.The disposition of proliferating liver cells

as an outer mantle around islets was sug-gested to be supportive evidence for theexistence of pancreatic hepatotrophins. Theregenerative pancreatic changes that havebeen demonstrated raise the possibility ofan hepatic pancreaticotrophin.

Faecal chymotrypsin accurately predictsexocrine pancreatic function in infants andchildren

G A BROWN. D SULE. J W PUNrIS, ANt) I W

BOOi iT (Institute of Child Health, Universityof Birmingham, Birmingham) Faecalchymotrypsin measurements have beenproposed as an alternative to intraduodenaltests of pancreatic function but withinsufficiently rigorous validation. We havereassessed the test by simultaneousmeasurements of faecal and duodenalchymotrypsin in 30 children aged threeweeks to 13 years with exocrine pancreaticfunction ranging from nil (five), throughpartial insufficiency (one), to the physio-logical range (24). Mean faecal chymo-trypsin concentration was determined, foreach child, from three random stoolscollected on separate days within 72 hoursof a conventional pancreozymin stimulationtest. In the 25 children with measurableduodenal enzyme, the mean faecal chymo-trypsin concentration was significantlypositively correlated with both apparentchymotrypsin secretion rate and mean duo-denal concentration (Spearman's rankcorrelation coefficients of 0-63 and 0(45,p<(-(X)l and <0-01 respectively). The fivechildren with undetectable duodenalchymotrypsin had mean faecal concentra-tions of only 3- 10% of the lower limit of thereference range 120( Rg/l(X) g stool). In thechild with partial insufficiency it was 33% ofthe lower limit. All 24 children with normalfunction were within the reference range. In21 neonates with suspected meconium ileus,faecal chymotrypsin was less than 19 sg/gstool in all 16 babies subsequently shown tohave cystic fibrosis. Faecal chymotrypsin isa rapid, simple. cheap. readily repeated,non-invasive test which should be donebefore contemplating intraduodenaltesting.

A 14(00


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Cytological diagnosis of hepatobiliary andpancreatic malignancy

I S BENJAMIN. T KRAUSZ. P t)OMIZIO, S

LAZZARA, AND 1. It HiUMGART (HepatobiliarsvSurgery Unit and Department of Histopath-ology, Hammersmith Hospital, Lonidon)During a six year period 162 patients sus-pected of malignancy of liver, bile ducts orpancreas were examined cytologically. Onehundred and twenty patients had provenmalignancy. 41 benign lesions, and oneremains uncertain. One hundred and ninetysix specimens were examined ( 118 pre-operative and 78 intra-operative). Malig-nancy was diagnosed correctly in 81/118cases (68-6%). Fine needle aspiration(FNA) was more sensitive (80(% true posi-tive) than exfoliate cytology (40)%/). Therewere no false positive results, but onesuspicious result in a patient whose bile ductstricture remains of uncertain nature. Fineneedle aspiration of biliary lesions had thehighest sensitivity both pre- (77% ) andintra-operatively (95%/O). The overall pre-dictive value of a positive result was 11)00,,(98-8%Yo if the unproven case is included) butthe predictive value of a negative result wasonly 54%. Fine needle aspiration cytologyhas proved valuable for pre-operative diag-nosis, especially in bile duct strictures, andmay also aid intra-operative diagnosis indifficult cases. Palliative intubation ofsuspected biliary tumours without tissueconfirmation should now rarely benecessary.

A novel enzyme linked lectin assay com-pared with RIA CA 19-9 as a serum test forpancreatic cancer

C K CuiING AND JONATilAN M RIIODES(University Dept oJ Medicinie aidt WaltoniHospitil, Liverpsool) We have recently des-cribed a novel serum marker for pancreaticcancer: a 35x 10" D glycoprotein thatexpresses galactose 1-3 NAc galactosamine(peanut PNA, lectin binding). This markerwhen present accounts for a high proportionof total serum peanut-lectin binding activity(PLBA). An enzyme-linked lectin assay(ELLA) has therefore been developed toquantify total serum glycoprotein gal 1-3gal NAc expression. This has been com-pared with RIA for CA 19-9 antigen.

Sera were diluted (1:20 t)00) in carbonatebuffer pH1 9-6 and coated (16 hrs at 4"C)onto micro-ELISA plates which were thenincubated in peroxidase-PNA (12 5 !tg/ml)after washing and quenching. Bound lectinswere estimated colorimetrically. Activity

was expressed in arbitrary units by compari-son with a standard curve of' a knownimatrker-positive serum. High PLBA (>1 ()unit Pl BA/ml) was tfound in 19/32 pancre-aitic cancer sera. 0)/18 normials. 8/20 othercancers. 0115 pancreatitis. 2/12 benigni and5/8 malignant obstructive jaundice. Elevenof 32 pancreatic catncer serai had normnliCA 19-9 levels but 4/11 of these contained> 1 uPLBA/ml. Overall sensitivity forPLBA assay was 59%' and specificity 79)'(96%/ if other cancers excluded). CA 19-9 inthis study had a sensitivity of 66(Yo butcombination of the two tests improved thisto 78%o.

This ELLA is highly specific for mralig-nant disease and seems likely to proveuseful as a test for pancreatic cancer. p(ss-ibly with CA 19-9 assay.

(GASiRO DUOFI.NAI 11 POSTi.RS

Gastric mucosal bleeding: what dose ofaspirin is safe?

P J PRICHARI), (G K KIIJCIIINGMAN, ANI) ( J

IIAWKFY (Department ofj Therapeutics,University Hospital, Nottinighaim) Lowdoses of aspirin are increasingly recoin-mended for the prevention of vasculardisease. They atre assumed to he safe butthere is no evidence that this is so. We havetherefore measured aspirin induced micro-scopic gastric bleeding to determine thethreshold dose for gastric injury in humans.Forty eight healthy volunteers were studiedunder basal conditions and after takingaspirin 75 mg or 30)0 mg/day for five and 12days or aspirin 1800 mg/day for five days.Mucosal injury was quantified as micro-scopic bleeding intragastric washings aspir-ated via an orogastric tube two hours afterthe final dose. Consumption of aspirin forfive days increcased bleeding from ba'sailvalues )-68()l/l() mins (95%/ confidencelimits 0(42-1 12 f[l/l() minsr to 1-32 (1 (1-2-31) d/10() mins (aspirin 75 mg, p<()-()l) to2-38 (1-7-4-2) d /10 mins (aspirin 300 mg,p<0.0l) and to 8X11 (5-36-1228) d/1t)mins (aspirin 1800 mg, p<0.0 1). Values at12 days with aspirin 75 mg (I1-27 (0-59-2-72)[d/l0) mins). and aspirin 300 mg (2-54(2-34- 995) fd/l() mins) were not signific-antly differcnt from fivc day values. Thebleeding caused by aspirin was dosedependent (p<().t)5) and implies a thres-hold for mucosal injury of about 30 mg/daiy.Thus even the lowest therapeutic doses ofaspirin increased the risk of gastric mucosalbleeding in humans.

Aspirin and bleeding peptic ulcerP J PRI( IiARi)D K W SOMIRVI Ii.G(FAUL KNI R,AND) M J S ILANGMAN (Department olf7hlera-e(lic(.s, Universit l Ho.spital Nottinlghamil,

Nottintg,alimii) Aspirin intaike hais beenrelated to uppcr ga.strointestinall bleeding,but the strength of this association hasrenmaiined unclear. Wc haive rcported thatothcr NSAIDs aire strontgly associated withbiccdilg peptic ulcer in the eldcrly. Wchave analysed the daital from this earlierstudy with rcspect to aspirin intakc.Two hundred and thirty of "29) paitients

cged ¢>6(0 idmitted with biccding pcpticulcer between 1983aXnd 1985 wercqucstioned. Two hundrcd anld thirtyhospital controls and 2(07 of 230 communitycontrols consented to qucstioning. Paiticntswith hiccding pcptic ulcer were more thaintwicc ais likely to bc taiking aispirin thanhospital or community controls with relai-tive risks (95%O confidencc limits) of 22(1-4-3-6) and 3-3 (1-8-6-2) rcspcctivcly(p<()-()()l McNemair's test). Thesc rcsultschianged little atfter cxclusion of concurrcnttakcrs of other NSAIDs aind taikers ofaspirin lor indigestion. Other ainilgesicintakc did not viary significantly betwcencases and controls. Both bleeding giastricaind duodenal ulccr were significantlyassociated with increatsed aspirin intatkc.Rcspective relative risks being 3(0 (1.3-7.4)and 3 3 (1.4-8 )) (community controls).The aittributaible risk lor atspirin and otherNSAIDs in bleeding peptic ulccr was 35%(28-42%) (community controls).

Aspirin intaike is significantly associatedwith biceding gastric aind duodenali ulccr inthe elderly. Risks appear to diffcr littic fromothcr NSAIDs.

Effect of simulated upper gastrointestinal(GI) haemorrhage on gastric acid secretionand G1 hormones

(G M FUILt ARiON, F. J BOYD), G P CREAN, K

BUCHANAN, ANt) K 1i 1t MCCOiLi (UniversityDept Medicine, Western Infirmairy atidGastroititestinIal Centre, Southernt GenteralHospital, Glasgow) It is surprising that themajority of acute upper GI bleeds stopspontaneously despite the aidverse aicidicenvironment for clot lormation. We haveexiamined the effect of simulated duodenalbleeding on acid secretion and G1 hormonesin seven healthy volunteers.

Gastric secretion was stimulated with ivpentagastrin infusion ((1.25 ug/kg/h). After4x 15 min collections of gastric juice 41) mlblood was venesected and infused directlyinto the duodenum before coatgulating

A 14()1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


every five minutes for 20 minutes. Further4x 15 min gastric collections were obtainedafter commencement of blood infusions.Corrections were made for pyloric lossesand duodenal reflux and tany duodcno-gastric reflux of blood was quantitated.Plasma concentrations of GI hormoneswere monitored. The effect of duodenalinfusion of egg-white which has a similarprotein and carbohydrate content to bloodwas also studied.

After commencement of blood infusions,gastric acid output (mmol/h) fell to 21 4±

3-7 (mean+SEM) compared to 29.8±3-3over the preceding hour (p<0-02). Mean (Yoinhibition was 30-4%0 (range 17-67). Thiswas due to reduction in volume and (H')and was not explained by reflux of blood or

duodenal juice or incomplete recoveries.

Plasma gastric inhibitory peptide (GIP)concentration (pmol/1) increased from32±2 (mean±SEM) to 66±9 followingblood infusion. Egg white did not affect acidsecretion or GIP release.

This inhibition ot gastric acid secretion byintraluminal blood may be mediated by GIPand represent a physiological response tofacilitate hacmostasis.

Intragastric fibrinolysis in bleeding pepticulcer disease

K 1. W1iFAIihY, VALERIE A POXON, P W )YKES,

ANI) M R B Ki.iAiiiY (Departtnent of Castro-entierologs, Genieral Hospital, Birmingham)Bleeding peptic ulcer disease is a common

problem with mortality rates of over 1)°/., in

many series. Little is known, however,about the aetiology of bleeding or rebleed-ing in these patients. We have studied therole of the excessive fibrinolytic activity ingastric juice, due to duodenogastric Trypsinreflux. Gastric juice was taken from 32patients presenting with upper GI haemorr-hage (DU-11, GU-9) and 31 controlpatients (DU-10, GU-8). Fibrinolyticactivity of aspirated gastric j uicC was

assessed using fibrin plates. and Trypsinassayed using a colourimetric method.Fibrinolytic activity was found to be presentin 20 patitents (63°/0) with upper GI bleed-ing, and only four (1 3%) of controls(p<.)()()5). Trypsin was detected in all

gastric juice samples showing fibrinolysis.Fibrinolytic activity was present in 22 out of29 samples with pl->4. and only two of 34 atlower p11 (p<()f()()5). Fibrinolytic activitywas scen in six of 11 patients with bleedingDU (55%), but only one of 10 control DUpatients ( 10% ) (p<().025). There wias no

significant differcnce in librinolytic activity

with bleeding or non-bleeding gastriculcers.We conclude that patients with upper GI

haemorrhage show increased fibrinolyticactivity in gastric juice, compared with non-

bleeding controls.

The effect of submucosal adrenaline on

blood loss from standard bleeding ulcers

S C (iiUNG, J W C LEUNG, M (iAILVINA, AND

w LEE (D)epartnments of'.SurgerY, Medicineant1d Anaesthesia, Prince of Wales Hospital.The Chiniese Universits oJf Ilontg Konig,Hontg Konig) To investigiate the possible use

of endoscopic adrencaline injection to con-

trol haemorrhage for bleeding ulcers, we

studied the effect of submucosal adrenalineinjection on blood loss in a standard bleed-ing ulcer model in dogs.Dogs weighing 20-25 kg were ainaesthe-

tised using thiopentone and anaesthesia was

m-aintained with halothane aind oxygen.

L'aparotomy was performed and thestomalch opened via an anterior gastro-stomy. Submucosail injections of 3 ml of1/10000 adrenaline or 10% sodium meta-hisulphite (carrier substance for .adrenialine)was made in the body of the stomach.Standard bleeding ulcers were made in theinjected area with the Quinton ulcer-maker. Ulcers miade in non-inijected areas

served as controls. The blood loss in the firstthree minutes was collected by inverting theulcer over a small dish and measured by

weighing.

Mean blood loss from control ulcers(n=25) in the first three minutes was

7.0±5.5 ml (SD), injection of sodium meta-bisulphite (n= 15) ciaused no significaintchange (6-3±4 4 ml) whercas adrenaline(n= 15) caused a significant decrease in theblood loss (2-9± 1.9 ml, p<0-05). Con-clusion: Submucosal adrcnailine calused a

decrease in the blood loss from artefactualulcers. Endoscopic adrenaline injectionmay have a role to play in endoscopichaemostasis.

Randomised controlled assessment of endo-scopic adrenaline injection for activelybleeding ulcers

J W LEUNG., C S CHUNG, R 1 STiT1L, AND

T J CROFTS (Combined Endoscopv Uniit,Prince of Wtales Hospital, The ChineseUniver.sitv of Hong Kong, Hong Kong)Sixty eight patients with actively bleeding(spurting or oozing) ulcers at emergencyendoscopy were randomised to receive

endoscopic adrenaline injection (aliquots of0-5 ml of 1/10000 adrenaline up to a total of10 ml) and no cndoscopic treatment. Theinjection group compriscd 34 paiticnts(M=26. F=8; age 22-85. mcan=51 years).nine haid a spurter. In the control group (34patients, M=26, F=8; age 17-86. mean=56years), 10 had a spurter. After endoscopyboth groups werc given intravenous H2blockers. Endoscopy was performcd 24hours later and injection repeated if neces-sary in the injection group. Strict criteria foremergency surgery werc adhered to in bothgroups: (1) haiemodynamic instabilitydespite 4 units of blood, (2) totail trans-fusion of eight units, or (3) rebleeding:haeematemesis/red aIspirate with tachycardiland/or hypotension. Initial haemostasis wasaichieved in 100(% of injection group. Theinjection group required less blood trans-fusion (3X8±2-8 units v 5-9±5-4 units,p<0.05) and less emergcncy surgery (5 v 14,p<0(02) than the controls. Iherc was nodifterence in mortality (3 v 2).

Endoscopic adrenaline injectionl is effec-tive in stopping active ulcer bleeding. Itsignificantly decreases transfusion and needfor emergency surgery.

Early blood transfusion promotes recurrentgastrointestinal haemorrhage

S 1) Bi AIR, R M GREENiIALGiI (INTRODUCED BY

R A PARKIN) (Department.s of Slirgers' tn(dGastroenterology, (Charing Cross a1(id We.s-inzister Medical School, Lotndoti) It hasbeen suggested that mortality caused byupper gastrointestinal haemorrhage may bereduced by restricting blood transfusion.We have assessed whether this is due to ananticoagulant effect in a prospectiverandomised trial.One hundred patienits with severe, acute

gastrointestinal haeemorrhage were random-ised to receive either at least two units ofblood during the first 24 hours of admission,or no blood unless their haemaglobin wasless than 8 g/dl or they were shocked. Minorbleeds and varices were excluded. As hyper-coagulation cannot be measured usingconventionial coagulation tests, fresh wholeblood coagulation was measured by theBiobridge Impedance Clotting Time (ICT)and the results expressed as mean ±SEM.The ICT on admission for the transfusion

group (n=50) was 44+±0)3 mins (normalrange 8 12 mins). This hypercoagulablestate was partially reversed by eairly bloodtriansfusion to 64±t)-3 mins at 24 hours(p<0(001). The 50 allocated to receive noblood had ai similar ICT on admission of

A 14()2


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


4.7+±04 mins but the hypercoagulable statewas maintained with ICT at 24 hours of4-3+±0.4 mins. Only two patients not trans-fused rebled compared to 15 in the earlytransfusion group (p<0)-001). Five patientsdied, and they were all in the early trans-fusion group.These findings show there is a hyper-

coagulable response to haemorrhage whichis partially reversed by blood transfusionleading to rebleeding.

Laser doppler assessment of human gastricblood flow

P M Al LEN, I CIIiESNER, K WHEAIL EY, AND M

GOLDMAN (Department.s of Surgery anidMedicine, University of Birmingham,Birmingham) Necropsy injection studieshave shown that the ulcer prone areas of thestomach and duodenum have fewermucosal arterioles, but physiologicalevidence of poorer perfusion is lacking.We have studied 19 patients, median age

7t) (range 27-83) using the non-invasiveLaser Doppler (LD) technique. Fluxmeasurements were made during gastro-scopy under diazepam sedation, using thePeriflux Pf2 instrument. Readings weremade at 9 sites: distal oesophagus, proxi-mal, mid- and distal stomach on greater andlesser curves, and pre- and post-pylorus.

In all cases, the stomach was macro-scopically normal. Flux varied from 40±28V (mean±SD arbitrary units of flow) in theantrum to a maximum of 92+35 V on thelesser curve. In the proximal stomach, themean value at each site exceeded 70) Vwhereas none of the five distal gastric siteshad a mean value above 56 V. Overall fluxin the proximal stomach at 77.5±39.8 V wassignificantly greater than that of 48X5±31-2 V in the antrum and pylorus (p<0-00(Student's t test.

This in vivo study demonstrates markedregionality of blood flow in the stomach andduodenum, being least in the ulcer-proneareas.

Prospective study of the effect of partialgastrectomy with Billroth-II and Roux-en-Yanastomosis on serum pepsinogens

J B M J JANSEN, P N M A RIEFU, I BIUMOND, If J M

JOOSTEN, AND C B H W LAMERS (Dept Gastro-enterology and Hepatology, UniversityHospital Leiden and Dept Surgery, CWZ,Nijmegen, The Netherlands) Subtotalgastrectomy is usually followed by mucosalatrophy of the gastric remnant. This

atrophy may bc caused by reflux gastritis orto a reduced trophic effect of gastrin on thegastric body mucosa. In order to differenti-ate between these mechanisms, we prospec-tivelv studied serum concentrations ofpepsinogcns A (PgA) and C (PgC) beforeand 10 days, 1/ year, and one year aftersubtotal gastrectomy in 22 patients, 11 ofthem randomly selected for Billroth 11 and11 for Roux-en-Y anastomosis. Serum PgAand PgC levels were measured by radio-immunoassay.

Fasting bile acid reflux was increasedfrom 10±0)5 [tmol/h before to 121*4±75.8[tmol/h after Billroth 11 resection (p=0-01),but was reduced by the Roux-en-Y pro-cedure from 0(6+±04 stmol/h to 01 ±01-1stmol/h (p=0-02). Serum PgA showed aprogressive reduction with time, indepen-dent of the type of surgery; 115+9 tsg/lbefore and 87±13 [tg/l 1i) days (p<0().5),75+±1) tg/l 112 year (p<0-05). aind 39+6[tg/l one year (p<)-()l) after Billroth 11-resection, and 95+8 ftg/l before and 67±8tg/l It) days (p<0)-05), 53±8 (tg/l '12 year(p<0(05), and 20)±2 ,g/l one year (p<.(Ol)after Roux-en-Y surgery. There were nosignificant differences between the twogroups of patients. Serum PgC levelsremained unchanged and the serum PgA/PgC ratio was reduced in parallel with thereductions in serum PgA.

This study reveals a progressive decreaisein serum PgA after gastrectomy, indicatingprogressive mucosal atrophy of the gastricremnant. As this decrease in serum PgAcould not be precluded by prevention ofenterogastric reflux by Roux-en-Y anaisto-mosis, we suggest that this atrophy resultsfrom a reduced trophic action of gastrinbecause of the removal of the antrum.

Diagnosis of occult intra-abdominal sepsiswith "'Indium labelled leurocyte scanning(ILS) - early versus late scanning

SARAII CIlESiLYN-CURTIS, C t1 BARBEIR, M C

AlDi)RIDGE, I) A CUNNINGHAM, ANI) 1i A F

DUDLEY (Academnic Surgical Unlit andDepartment of Radiology, St Mary'.sHospital, Londoni) Intra-abdominatl scpsisremains a major causc of morbidity andmortality in surgical patients. ''llIndiumlabelled leucocytc scanning is an establishedmethod of diagnosing occult sepsis but itsmajor disadvantage compared with ultra-sound or computed tomography is thesignificant delay beforc a result is obtained.This has been reduced by faster labellingmethods but pure granulocyte preparationsare advocated when the white cell count

(WCC) is normal (< 12 x 1 /1). The imr1port-ance of the ealrly scani (three to lour hours)and the use ol mIixed leucocyte preparationsin patients without a ICLucocVtosis has notbcci established. We have evilulted 101scans in 81 patienits with suspected intra-abdominal sepsis to assess (a) the diagnosticaccuracy of the early scan (three to livehours) and (h) the relationship ol the bloodlcucocvtc counlit to thle scilsitivitv of I [S.lThe positivity of Il S was conlfirimied hvoperation or spontaileOLiS dischairge of ptusind negativity by recovery of the paitienitwithout further surgery.

In only four of 73 patients with hoth earlyand latte scans wais interpretation altered hythe late scan. Twenty two of 23 scans inpaitients with a normail WC C were coni-firmed cliniclily but with a WCC of>20x 10()/1 five of 30 scans were equivocal1 orgatve false results.

It is concludcd that the early scan isdiaignostic in most caises but 1f It Is neICtgivcor equivocal late scanning is recotiieidided.A normal WBC does not exclude the use ofthis investigation and special preparationmethods are not needed. Very high WB3Cscan be paradoxically -associated with diag-nostically inaIccurate scans.

Impaired gastric adaptive relaxation inpatients with postvagotomy diarrhoea

M N HARI I.IY ANI) ( R MA(Ksit (UnIiI'-rSiltVDepartment of Sutrgers', Royal LiverpoolHospital, Liverpool) Gaistric idaIptivCrelaxation (GAR) is said to he impairedifter truncal vagotomy and drainage(TV+ D). Ioss of GAR after vagotomy maycontribute to the excessively rapipd gastricemptying exhibited by patients with post-vagotomy diarrhoea. We studied GAR in1) healthy subjects (mean age 29 years,range 22-4(), and 17 paitients. six mouiths to17 years after TV+D (mean age 54 years,range 39-65 years) of whom six patients hadpersistent postvagotomy diarrhoea.

Fasted subjects were intubated with aRyle's tube with a flaiccid plastic bag (800ml) attached and containing a pressuremicrotransducer. Gastric corpus-funduspressure waIs recorded during distension olthe bag with 460)±30 ml (mean+SD) ot airover 30 sec. Pressure indices (P1) werederived (areat under curve: mean of fourreadings) aind compared.

PI cm H1O (mean+SD) were: Controls11 7+3(0 v TV+D (without diarrhoea)15 4+2-2 v TV+D (with diarrhoea)21l2+3 6 (p<0O01 ANOVA).

This confirms GAR is significantly

A 140)3


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Societ7 of Gastroetiterology

impaired after TV+D and the degree ofimpairment is significantly greater inpatients with diarrhoea thain without.

Highly selective vagotomy (HSV) in thetreatment of acute complications of pepticulcer

M RM(EiRS, I) JOIINSTON, J N PRIMROSFI, I) CWARI), R 1. BL ACKFETI, ANI) M J MCMAIION

(University' Depa(rtment of Sutrger_y, TheG,enerdl Infirma(ry, Lee(ds) Sincc the adventof the H2-reccptor antagonists, elcctivesurgery for pcptic ulccr (PU) has becomelcss common. The complications of PU.howevcr, continue unabated.

Highly sccctivc vagotomy is the bestcicctive surgical treatment for PU but whatproportion of paticnts with acutc complica-tions can be trcatcd with H:SV, and howeftectivc is it in curing the ulcer'? Thcsc arevicious ulccrs and onc might expect theincidenicc of rccurrcnt ulccr (RU) to bchigher aftcr HSV for complicated PU thanafter clective IISV for PU.

Since 1969. 985 paticnts havc undcrgoncH1SV in this department of whom 127 (13%)underwent [ISV for acutc complications ofPU (perforation - 70, haemorrhage - 57).Highly sclective vagotomy was used in 33%!of pcrforated PU's and 45% of bleedingPU's. Overall mortality was I ("o (perforatedPU- 0%, bleeding PU - 2%). Incidence ofRU was 13% for perforated PU's, 2% forbleeding PU's (overall 8X/) - no differentfrom electively treated PU's. Visick grad-ings were not significantly different frompatients with PU's treated with electiveHSV. In conclusion, HSV can be used inpatients with acute complications of PU,and with careful case selection it carries alow mortality, low risk of RU and gives agood quality of life.

Clinical results 10 years after proximalgastric vagotomy

C MULLER, R IE'(IIMANN, P Vi-RREE'lI, B

IIUSEMANN, 1. FIEDI)ER, ANI) B FNGLKE (IN'TRO-1)U1CE D) BY .1 11 BARON) (Dept of SurgerY,Kanitonsspital, CH-4031 Basel, Switzerland)To assess long term results of proximalgastric vagotomy (PGV) sufficient numbersof patients actually followed-up for 10 yearsor more are needed. We report on a pro-spectivc multiccntre trial on PGV initiatedin 1974.Betwcen Jainuairy 1974 and April 1975 717

patients underwent PGIV: 524 for duodenal

ulcer (DU), 94 for pyloric or prepyloriculccr (PU/PPU), 71 for gastric ulcer type I(GU) and 28 for combined ulcers (GDU).Complete follow Lip wis achieved in 85-4`%,55"` of which had routine cndoscopy.len ycirs' total recurrenice raite (includ-

ing aisymptomatic recurrences) was 20'4%for DU 32% for PU/PPU, 198) for GUand 317"% for GDU. Complcte vagotomyas aissessed at operation by at modifiedBurge-Test significantly lowered the recur-rcncc rfatc for DU to 13-6%). In PU/PPU adrainage procedure significantly improvedthe outcomc ( 16'6%° with versus 39%, with-out drainage). At 1) years Visick gralde Iiand 1I wias found in 88%. of the DU, 84% ofthe PU/PPU, 73%. of the GU and 82', ofthe GDU.The ovcrall symptomatic result 1t) years

aifter PGV is good in all ulcer types. Thechaincc of cure is 80%O for DU and surpris-ingly for GU. but unsaitisfactory in PU/PPUand GDU. 'Ihc outcomc is significaintlyimproved by an intraoperativc complctc-ncss test in DU alnd by a drainage proccdurein PU/PPU.

Asbestos: a possible aetiological cause ofgastric cancer?

P W J IIOU(II'ON, I STEIWART, P iHiAP, N J MCC

MORTINSEN, ANt) R C N WIllTlIAMSON (Depart-ment of Surgery, Bristol Royal Infirmaryand Department of Anatomy, University ofBristol, Bristol) There is some epidemio-logical evidence to suggest that workersexposed to asbestos have an excessivemortality from gastric cancer and it has beenpostulated that the high incidence of gastriccancer in Japan may be due to contamina-tion of their polished rice by asbestos.

Scainning electron microscopy was per-formed on specimens of stomach takenfrom six patients with histologically provengastric cancer and on three patients under-going gastric resection for benign disease.In five of the six patients with gastric cancerunknown fibres were found in the tumourtissue and in one patient in the premalignantmucosa surrounding them. Simultaneousradiographic microanalysis of these fibreshas shown them to contain silicon, mag-nesium, calcium and iron and their molarratios strongly mimic the known x-ray ele-mental profile of tremolite asbestos. Nofibres have been identified in any of thecontrol tissue.These findings suggest that mineral fibres

might play an important role in the patho-genesis of gastric cancer.

A randomised, controlled study of adjuvantTamoxifen in the treatment of gastriccarcinoma

J D IIARRISON, I) 1. MORRIS, 0 Fils, -I JONF.S,AND) I JACK(SON (Department of Surgerx'vUniiversitY Hospital, Nottinighiamlte) Toinvestigate the effect of Tarmoxilen inpaitients with gastric cancer, 100 patientswere raindomilised after stratilication totreaitment or control groups. 'Ihe treatmentgroup received 4t) mg laimoxifeni dailly aindthe patients were followed up tintil deaith.The mediain survival in men was 25-5 weeksin the control group and 15 5 weeks in thetreaitment group. whilst armongst thewomen, the mediian survivail in both groupswats 25 weeks. Fo investigiate the efiect ofoestrogen receptors on prognosis thesepaticilts tumours had oestrogen receptorstatus measured. Ihe overall cumulativesurvival in patients who were oestrogenreceptor negative (ER -) was significantlybetter thain in those piatients whose tumoursexpressed oestrogcii receptors (ER+1),particulatrly in the femalle group. Themcdian survival in ER+ women wa1s 185weeks compared to 33 weeks in the ER-women (Logranik test p<0)(01).() Themedian survival in men was 20 weeks lorER+ patients and 22 weeks for ER-Thcse studies show thatt the presence oloestrogen receptors in gastric carcinoman isan indicator of poor prognosis in women,and suggest that sex hormone manipulationin men may have a role in the trea.tment ofgastric carcinoma.

C-myc oncogene product expression inbenign and malignant gastric epithelia

W 11 AlTlUM, K M NEWBOTLT), F MACD[)ONAiL[),B RUSSETLL , AND i J SI'OKES (IN TROD)UCTH) BY

M R B KEi(GiLEY) (Slurgical Immnunology Unitand Department of Pathology,, QueenElizabeth Hosspital, Edgbaston, Birming-ham) Amplification of the C-myc oncogenehas been described in gastric cancer. In thisstudy the expression of the oncogene pro-duct, p62 C-myc, has been assessed in bothbenign and malignant gastric cpitheliai todetermine whether C-myc is important ingastric neoplasia. Histological sections of arange of benign and malignant epitheliahave been examined in an immunohisto-chemical assay using a monoclonal antibodyto p62 C-myc. In the benign tissue, stainingwas cytoplasmic. All tissue types showedsome staining but gastritis was morecommonly positive than normal epithelium.This was particularly marked in atrophic

A 14()4


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Societv of Gastroetnterology

gastritis showing type 2b intestinal meta-plasia. There was no diffterence betweenactive and quiescent chronic suLpcrficialgastritis suggesting that staininig was notsimply demonstrating proliferiation. In asignificant proportion staining was limitedto the tips of the mucosal fcolds, sites notusually considered to be actively dividing.This was most marked in type 2b intestinalmetaplasia. In tumours, less than 40(% ofspecimens stained positively. When presentstaining was unreliated to differentiation.although there was a suggestion ofincreased positivity in intestinal typelesions. C-myc may he important in theprogression to gastric malignancy and mayregress oncc malignant change hasoccurred.

Use of CA-50 in the differential diagnosis ofbenign and malignant diseases of thestomach and oesophagus

S B KFl-llY, M J HEiRSIHMAN, N A IIABIB, R C N

WILt.IAMSON, J SPE.NCER, AND C Bt WOOl)(University' Departmenit o)f Surgerv, BristolRoyal In/irmarv, Bristol acnd Department ofSurgery, Royal Postgraduiate MedicalSchool, London) I his study investigatedthe role of the tumour marker CA-5()(carcinoma associated antigen) in thedifferential diagnosis of benign andmalignant diseases of the stomach andoesophagus. Serum was collected from 5()controls, 19 with benign oesophagogastricdisease, 24 with gastric carcinoma and 21with oesophageal carcinoma. A radio-immunoassay (RIA) was used to detectCA-5) in the serum and a level of 17units/ml was used as a cut-off betweenbenign and malignant disease. All 5S)normal sublects and 18 of 19 (95'Y1) withbenign oesophagogastric disease had CA-5)levels below 17 units/ml. In the cancergroups, 18 or 24 (75°/) with gastriccarcinomai and 15 of 21 (71"%) with oeso-phageal carcinoma had CA-5( levels above17 units/ml. Therefore, the sensitivity is73"% (33 of 45) and the specificity is 100%,(5(t of 50) and 95% (18 of 19) for the controland benign groups respectively. In patientswith tumours and values above 17 units/ml,the mean concentration of serum CA-50was 65+33 (range 31-132) for the gastriccarcinomas and 58±30 (range 19-116) forthe oesophageal carcinomas. These datasuggest that the CA-50 RIA test could be ofuse in the differential diagnosis of benignand malignant diseases of the stomach andoesophagus.

Analysis of dyspeptic symptoms to deter-mine rick of gastric cancer and priority forendoscopy

1 M (CHISIHIOI.M, B.1 UNWIN, A ( MOR(F AN. AND)( R (ill- 1S (.St JaInleS'S Hosp)iatl, Leed.s and(lA ire(lale Hospital. Keighley, Yorks) Scrccn-ing for early gastric c Gncer(EGC) by gastro-scopy in all middle-aged piatients with a twowcek history of dyspepsia has recently beciadvocatted. Unfortunately a 2-4A" yield of'ciancer with at 98% false positivc rate arose.We hiave attempted to further define risk ofcancer and there'ore priority t'or endoscopyin dyspeptic patients by analysis of theirsymptoms.The relative incidence of symptonis in 10(1

paitienits with glastric cancer (2(1 EGC), 164with benign upper gastrointestinal condi-tions and also in 30)0 middle-aged randomilpopulation controls werc rccorded. AsimFple scoring index wias then produced byapplying the Log Likelihood Ratio statisticto these symptom frequenicics to prcdict'high risk' of cancer (defined ats probability>0-05). Eighty one cancer patients wcrcabove this value with only a 6% fialsepositive rate, giving a 74% yicld for cancer.

[o determine its clinical significaincc, thescoring indcx was recordcd in 300 consecu-tive dyspeptic patients and was compiaredwith a cliniciains evaluation of cancer risk.Fifteen of 16 gastric cancers (4/4 EGC) werepredicted by the index but only 9/16 werefelt to havc canccr by the clinician. Theseiisitivity, specificity and yield for ciincerby the index compares well with theclinician's.A simple risk index for gastric ciincer

using symptom analysis cain predict cancerrisk as accurately as a clinician, with a highyield per endoscopy, and maiy therefore bc atuseful atid in determining priority for endo-scopy when screcning for EGC.

Ability of duodenum to maintain a neutralpH microclimate in response to acidchallenge in endoscopically normal andduodenal ulcer subjects

B J / DANESII, 0 STARK, J M RAWI iNGS, M 1.

ILUCAS, ANt) R I RUSSEI.II ((Ga.stroenterology'Unit, the Royal Infirmnarland DepartmentoJ Physiolo,gy, University oJ Glasgow,Glasgow) It has recently been shown in manthat gastroduodenal mucosal surface islined by a near-neutral pH mnicroclimatc.Wc measured duodenal mucosal surface pHon direct acid chcillenge in 11 endoscopicaillynormal subjects (controls) and in 11

paiticnts with aictivc duodenal ulccration(DU). Using ai pH clectrode indirodLucedthrough an eiidoscope. iiucosail and luiriaiit.p1lI wais meaiistired in situ in the Itilltius,aintruni, duodenall Ceap aiid loop bcfore aiiidduring perfusioi with phosphaitc (p11 1 5)hulffer. 13efore icid perf'usioii, the duodenaliiiucosaIi pHl wats higher (p<0f-00l1) in DLJ(72±0( 1 ) thain in controls (6118+(1 ).D)espite comrlprablc 'Lfuiail lLnlllililil ieCidityin conitrols (24±0-.5) ndi DLJ (18±0 I),lurniiril p1 wais morc ailkialinc (p<()()I5) inthe duodeOnld loop in DU (78X()1 ) thlin inithe coiitrols (7.4±()02). Acid perfusioiireduccd (p<()0()0l) mucosail p1H in thezintruLii, duodeciail cap aind loop in DU(ApH= 1 1+±04, 1 3±0-3, 1-1±0-2) but notin controls (Apl0=()2±0()1, ()0()7+0()()8,)- 1 +()-()7). During icicd perfusiosIn, MUCOSapH wats lowcr (p<()()l ) in the duodcnal calparnd loop in DU ptittenits (5.8±0(.3, 6(1±0 1than in the conitrols (6-8±011, 6(7±() 1).

[Ihe rnucosal alkalinity observed in DU,not mniiintnied diuriiig aicid chal clegc.suggests that riiucosal bicarboiiate secrctionIis ait ai maximium drive in DU and theCaIpaIcity to niutratlise fully ain aicid surgcseeriis limited.

Increased formation of leucotriene U.4(LTC4) in campylobacter pyloridis (CP) -associated gastritis

A AIIMIFI), 1) VNIRA, S K (CIRNS, H01lTON, M

Fii ZON, A POt' YI)ORUS, ANI) P R SAL1MON(Dept (Gastroe,nterolosg, MicrobiologY,Histopatholog/,, The Middlesex Ilospital,Lonldoni) Vhe prcsciicc of CP in the pgstricmucosa riay bc aissociited with inflmrni-am-tory cell infiltraltion aid colisequcilt alter-tion in l TC-4 aid prostaglaindini (PG) E.levels which could play ai proinflariimatoryrole.

Antral biopsies of 1 3 (yspeptic patientswere incubated ait 37"C in pre-wairmed/oxygenated Tyrode Solution (baisail recliesc)aind tramnsferred to sccoInd incubationrnediuni containing ioiophorc A23 1 875 Itg/ml (stiriulated relcase'). Superriatantcicosanoids were riieisured hv radiodimmunoaissiay.

Eight patients haid acute gstritis (intralCP prcsenit in atll) aind fivc haid normil aitralhistology (no evidencc of CP).

Baisial reiclsc (pg/mgww/2() min, macics+SEM) of LTC4 aind PGE, in norml(74±11.3 aind 1691±336.2) aind ihbnormilinucosai (94± 14-2 arnd 1824±331)3) weresimilalr. Stimulated release of l TC,1(147±21-1) wias signilficaiitly cnhainced corii-

A 1 405


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


pared to basal levels in gastritis (p<(0)5).Stimulated LTC4 in abnormal mucosa wassignificantly greater than normal mucosa(81±9 2, p<0)05). Basal and stimulatedPGE, levels in normal versus abnormalmucosa were similar.These results suggest that LTC4 forma-

tion may be of greater importance thanPGE. synthesis in CP associated acutegastritis probably reflecting inflammaitorycell infiltrate.

Comparative study of 4-hour rapid ureasetest (RUT) to culture and histology for thedetection of Campylobacter pyloridis (CP)in gastric and duodenal biopsies

D VAIRA, J iJOLfON, M FAILZON, A AIIMED, S R

CAIRNS, AND P R SALIMON (Dept of Gastro-enterology, Microbiology, Histopathology,Middlesex Hospital, London) The clinicalsignificance of CP in gastroduodenal biop-sies is uncertain. The aim of this study wasto evaluate the sensitivity and specificity ofthe four-hour RUT to histology and CPculture in the detection of CP in gastro-duodenal biopsies.

Gastric and duodenal biopsies from 24dyspeptic patients were homogeinised in0-9% saline and immersed in 2°/O ureabroth.

Biopsies from 16 patients (66.6%0)showed evidence of gastritis and/or duo-denitis, despite 5 showing no macroscopicalabnormality. All these 16 patients showedCP on histology, indicating a high correla-tion between the presence of CP and acutegastric and/or duodenal inflammation.

Determination of the sensitivity for RUTto histology and culture was 81l3'Y0, 56.3°/0and specificity was 1(N)% and 100(°/respectively.

This study shows: (1) Even when nomacroscopic abnormality is found at endo-scopy of dyspeptic patients histology andfour-hour RUT may reveal antral CP. (2)The four-hour RUT can provide a rapid anduseful method for the assessment of acutegastritis and/or duodenitis in dyspepticpatients.

Smoking and pH-response to H2-receptorantagonists

C-C SCFHUERER-MAIY, 1 VARGA, H R KOEL1/,AND F HiALITER (Gastrointestinal Unit,University Hospital, Inselspital, Bern andGastrointestinal Unit, Triemnlispital, Zurich,Switzerland) Smoking has been shown toimpair the effect of H.-receptor antago-

nists. To evaluate onset and duration of H-receptor antagonist action in relation tosmoking habits, we tested the effect ofranitidine (RAN) and famotidine (FAM)under physiological conditions, usingambulatory pH-metry. Intragastric pH wasmeasured over 20 hours. Each of 18 heallthyvolunteers. 20-36 years. nine smokers andnine, age and sex matched non-smokers,received either 40 mg FAM, 300 mg RANor Placebo in a double blind, randomisedstudy as a single evening dose, 1800. Withboth drugs 20) hours acidity was markedlysuppressed. After FAM treatment meaninhibition in smokers was 42°/., in non-smokers 76°/o, with RAN 60%, and 67%respectively. When areas under the pH-curves from each individual were calculatedand treatment compared to placebo(=1()00%/,), with either drug response wassmaller in smokers than in non-smokers(FAM 153±21°., versus 214± 19'%, p<0(01,RAN 176±21/, versus 232±29°/, p<0.05)during the first four hours after drug intake.A similar effect was observed in the morn-ing period from 6-100t) (FAM 118±19%versus 20)6±19'(o, p<0(001, RAN 133+21% versus 207±31%S, p<0(02). Duringnight time there were no significant differ-ences.These results indicate that smoking

impairs onset and duration of response ofboth drugs tested.

Smoking delays gastric emptying of solids

G MILLER, C FARRIN(G[ON, V SMITH, M V

MERRICK, AND K R PAL MER (Departmnentls ofNuclear Medicine and the GastrointestinalUnit, Western General Hospital, Edin-burgh) The effects of smoking upon oeso-phagogastric transit were defined in eighthealthy volunteers by radionuclide scanningan ingested test meal. Oseophageal clear-ance was measured by examining threeareas of interest during a liquid swallow ofIn"' labelled water and during a semisolidswallow of Tc'"1 labelled omelette. Gastricemptying of the swallowed test meal wasthen mcasured by a standard double count-ing technique. Finally valsalva and posturalmanoeuvres werc used in an attempt toprovide gastro-oesophageal reflux (GOR).Each subject was studied under basal con-ditions, whilst continuously smoking andthen whilst chewing nicotine gum.

Oesophageal transit was unaffected bysmoking or gum (mean basal liquid transit7 5±1 (SEM) smoking 66± 1, gum 6 3±1sec). Gastric liquid emptying was cxponen-tial and unaffected by smoking or gum T'/2

20(6±2A3, 21 4±7-3, 13-6±2-1 mins).Gastric solid cmptying had two com-ponents, both of which were prolonged bysmoking but not gum. An initial lag phase(mean baisal, 184±2 7, smoking, 28+5-7(p<0(05). gum 18+3 3 mins) and a subse-quent linear emptying phase (mean basal1(07±0+13, smoking 0(85 ±)009 (p<0(05),gum 0(98 ±009o(Y/min). No episodes ofGOR were observed.

Gatstro-ocsophageail reflux is associatedwith delayed solid but not liquid emptyingaind our findings suggest a mechanism forthe observation that smoking exacerbatessymtoms in susceptible individuals.

Mucosal repair of rabbit duodenum: role ofan alkaline micro-environment

W FFill, P KARNER, S KIiMESCII, M STARLINGER,ANI) R SCHIIFSSEI (INTRODUCED BY A GARNER)(Surgical Dept 1, Univ ('linic of Vienna,A-1090 Vienna, Austria) The duodenalmucosa of the rabbit has the ability to repairitself rapidly after acid injury. In this studywe investigated the importance of nutrientbicarbonate and of the layer of necrotictissue (=alkaline micro-environment) forthe repair process after acid damage.

Mucosal sheets of rabbit duodenummounted in Ussing-chambers were exposedto 10 mM luminal acid for 10 min. Afterinjury the luminal saline solution was eitheradj usted to pH -=7 4 or pH1 =3 and thenutrient solution contained either 25 mMHCO- or HEPES. In some experimentsthe necrotic layer was removed by scraping(RNL). Tissucs were allowed to recover fivehours after injury. We measured: alkalinesecrction (AS), potential difference (PD).and performed histology.

Acid exposure caused a 67%/ drop in PD(from 2 9 to t)9 mV), a 28%M increase of AS(from 0(67 to 0-86 r&Equ/cm2- 1() min) anddamage of 950O of villi (n=48). After fivehours morphometry showed repair of 72%of villi (HCO -, pHl -=74), 52% (HCO3-,pfi =3), 69%X (HEPES, pH1 =7 4), 0%Yo(HEPES. pH _=3), 52%° (HCO -,pH1 =7-4, RNL) and 27% (HCO-pH, 3. RNL) respectively (n =8, echgroup).The absence of nutrient HCO- did not

impair the repair process at pH[ =7-4 but atpH, =3. RNL delayed repair at pH, -74but prevented this process almost com-pletely at pH1 =3.We conclude, that mucosal repair in the

duodenum is dependent on the presence ofthe necrotic layer and the availability ofnutrient HCO-.

A14()6


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Synthetic secretin secretolin (SS): anadequate substitute to GIH natural secretin(NS) for secretin provocation test inZollinger-Ellison syndrome (ZES)

M MIGNON,. IFLOUAER-IBLANC( D RIGAUD. P

RUS/NIFWSKI, T CHIEIVAIEIIR, 1 VAI1OT, J

VATIIR, E RENFI, AND M D)UET (Service 1tiGastroenterologie, Hopitial Bichat, Parisand Laboratoire de Radioitnmultnologie,Hopital Lariboisiere, Paris, France)Synthetic secretin secretolin is equipotentto NS on gastrin release in control andduodenal ulcer (DU) subjects (1987) butthe comparativc efficacy of SS for ZESdiagnosis was not assessed. As GIH NS willno longer be available, both Hoechst SS andNS effects on acid, gastrin secretion wereevaluated in 10 unoperated ZES (53± 13years) and 13 DU (43+ 16 years). Inrandomised sequences 3 CU/kg SS and NSwere infused for one hour after antisecre-tory drugs withdrawal for three days. Serumgastrin (SG, pg/ml), acid output (AO,mmol/h) were measured simultaneously:four SG samples and one hour AO wereobtained before and upon secretin infusion.Parametric or non paratmetric paired-dataanalysis was used according to results distri-bution. In ZES: basal AO (BAO) was25±11; AO rose more on SS (AO; 35-4+15) than on NS (AO: 31-+816); differenceBAO versus (v) AOSS and AOSS v AONS:p<)-05. Basal SG (BSG) was 219 (84-7813)and increased more on SS (SG: 950, 118-18769) than on NS (SG: 834, 92-12(012);difference BSG v SGSS and SGSS v SGNS:p<0-05. In DU: BAO was 9-7±6-5; AOdecreased more on SS (AO: 0-57+ 1-1) thanon NS (AO: 128± 1-25); difference BAO vAOSS or AONS: p<0)-)01. DifferenceAOSS v AONS: p<0(05. SG wasunchanged by both SS and NS. Thus com-pared to NS, SS induced higher inhibition ofAO in DU, higher increaise in AO aind SG inZES and can be confidently substituted toNS for ZES screening. *(M±SD)

GASTRO D)UOD[NAI POSrFERS III

Food is bad for ulcers

1) A JOHNSTON AND K G WORMSLEY (NinewellsHospital, Dundee) Inhibition of nocturnalgastric secretion is thought to be the mostimportant factor in determining the healingof duodenal ulcers. Overnight gastric secre-tion was measured after an oral dose of 300mg ranitidine at 22 00 h in nine patients with

duodenal ulcers which had failed to heal intwo months or more. In all patients, gastricsecretion was completely aibolished duringthe night, from which we conclude thatresistance of the ulcer to heialing is not dueto failure of nocturnal giastric inhibition.Eight patients with nonhealing duodenalulcers underwent two 24 hour studies ofgastric secretion. On one day, the patientsreceived 15t) or 30)0 mg ranitidine at 10)0( hand 220)( h and standard mealls were eatenat 080)X, 1300, and 1800 h. On the secondstudy day, patients were fasted throughoutthe 24 hours, except for the two doses ofranitidine, as previously. Gastric aiciditywas significantly less inhibited between14()0 and ()1 00 h when meals were eaten.Nocturnal (00 00-08t)(0 h) gastric secretionwas similarly inhibited on both study days.We conclude that food significantly inter-

feres with therapeutic gastric secretoryinhibition. It seems to us that thisphenomenon contributes to resistance ofsome ulcers to healing with ranitidine. Wehave therefore healed resistant ulcerswithin 14 days by withholding food andadministering nutrition parenterally.

Human antropyloroduodenal motor activityafter intraduodenal acid infusion

D D KERRIGAN, 1t A HOUGHTON, N W READ, ANDA G JOHNSON (Dept of Surgery and Sub-Deptof Human Gastrointestinal Physiology andNutrition, University of Sheffield, Sheffield)A manometric catheter was used to recordpressure activity in the antrum (three sites),pylorus and duodenum (four sites) in 12healthy fasted volunteers during intraduo-denal infusion of normal saline and isotonic)- I M HCI. Pyloric pressure was recordedusing a 4 cm long sleeve sensor positionedby measuring the transmucosal PD at eitherend of the sleeve. Intraluminal pH wasmeasured in the terminal antrum and at twopositions in the proximal duodenum. Salineand acid were infused into the duodenalbulb alternately for consecutive 30 minuteperiods for up to three hours at both I and2 ml min. During acid infusion there was (1)a significant reduction in the frequency ofantral pressure waves [acid, 8h (0-54h ')median (range); saline, 87h (10-272h '):p<0.011, (2) a marked increase in pressurewaves confined to the pylorus [acid, 18h-X(6-128h l);saline, 2h '(0-14h '): p<0)-01,(3) a reduction of coordinated pressurewaves involving the duodenum [acid, 13h(0-50h '); saline, 45h (24-70h '):p<0001. These changes were reversedduring subsequent saline infusion but

returned during the next acid infusion. 1'husour data indicates that acidification of theduodenum may slow gastric emptying bycausing reproducible reversible inhibitionof antral pressure activity, increasingpyloric resistance and reducing duodenalcoordination.

Reproducibility of 24-hour studies of intra-gastric acidity, nocturnal volume, acid andpepsin secretion by the aspiration method.

M DFAKIN AND J G WIllIAMS (Department oJGastroenterology, Royal Naval HospitialHaslar, Portsmouth, Hants) Whilst 24 hourstudy techniques have been widely used inthe evaluation of antisecretory drugs therehas been no previous formal assessment ofreproducibility.

Eight volunteers with duodenal ulcers inremission were studied during two 24 hourperiods using a size 10 French sump typenasogastric tube. Three standard meals of375 ml of clinifeed with one Oxo cube in 2()()ml of hot water were taken during each 24hour period at ()X8)(), 1300), and 18X)0 hrs.Gastric aspirates were taken for pH record-ing at 15 minutes after each meal for threehours and half hourly between meals.During the night the stomach was keptempty by a continuous, intermittentpositive pressure, aspiration pump.The intragastric pH profiles were virtu-

ally identical on both days in all eightsubjects. For each patient pH scores for the24 hour period varied by 6-8%". (0)-21%);median (range). Nocturnal volume, acidand pepsin secretion were more variable.Median percentage differences betweendays 1 and 2 were: acid output 36%, (2-661%); pepsin output 80X% (35-210%));pepsin concentration 41% (0-93%) andvolume output 50% (5-153%).

Intragastric pH is the most reproducibleand most convenient parameter for thestudy of a drug effect. Outputs are morevariable making measurements for an indi-vidual patient unreliable.

Effect of meal temperature on gastric empty-ing of liquids in man

W M SUN, 1 A HOUGHTON, N W REA[), D

GRUNDY, AND A G JOHNSON (Dept of Surgeryand Sub-Dept of Human GastrointestinalPhysiology and Nutrition, University ofSheffield, Sheffield) The existence ofthemoreceptors, responding to tempera-tures of 1()-1 2°C and 46-490C in the mucosaof the stomach and duodenum suggests the

A 14()7


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


possibility that the temperature of food mayinfluence the rate of gastric emptying. Intra-gastric temperature from two sites andgastric emptying rates were measured con-tinuously in six normal volunteers afteringestion of 400 ml radiolabelled orangejuice at either 50, 37, or 4°C. The highest(43±()±0(4°C, mean±SEM) and the lowest(2 1 *2+1 *9°C) intragastric temperaturesoccurred within a minute of ingestion, thenreturning to body temperature within 20minutes. Differences in intragastrictemperature between the warm and controldrink were significant up to six minutes andbetween the cold and control up to 10minutes after ingestion (p<0-05). Theinitial rate of gastric emptying of the colddrink was significantly slower than the con-trol; at 10 minutes 76+7% of the cold drinkand 59±4% (p<0.05) of the control drinkremained in the stomach. The difference inemptying rates between the cold and con-trol drinks were significantly correlated withdifferences in intragastric temperatures(r=0-98; p<0-(l). Initial emptying of thewarm drink also appeared slow, was notsignificantly different compared with con-trol. Thus our data indicated that mealtemperature effects the rate of gastricemptying when intragastric temperaturediffers from body temperature.

Validation of non-invasive assessment ofduodenogastric reflux (DGR)

D C WARD, D JOHNNS1ON, I P BELTON, R F i J

KING, AND M C J BARKER (Uniiversity Depart-ment of Surgery and the Department ofNuclear Medicine, The General Infirmairy,Leeds) Duodeno gastric reflux has traidition-ally been investigated by studying gastriccontent aspirated through a nasogastrictube (NGT). The NGT may itself, however,affect DGR. Although well counting ofgastric aspirate after administration of`Tc1'-EHIDA has been compared with bileacid concentration (BAC) in gastricaspirate, external gamma camera imagingafter'Tc'"-EHIDA has not been comparedwith traditional methods involving a NGT.We studied DGR both by ""Tc"'-EHIDA/

external imaging (DGR-HIDA) and byestimation of BAC in gastric aspirate(DGR-NGT) in the same 48 patients. Inseven, the reproductibility of DGR-HIDAwas studied by repeat imaging after 48-72hours. In a further seven subjects imagingwas repeated after 48-72 hours, but with anasogastric tube in situ.DGR-HIDA in the second test did not

differ significantly from DGR-HIDA in the

first test (p=NS, r=0-65). The presence ofNGT did not affect DGR significantly.DGR-HIDA was found to correlatesignificantly with DGR-NGT (p<()-()()l forboth peak and post prandial DGR). WhenDGR-HIDA and DGR-NGT weremcasured simultaneously correlation wasexcellent (r=O-88, p<O-0l).Thus "Tc'-EHIDA/external imaging

provides a convenient and reliable methodfor measuring duodenogastric reflux.

Value of the alkali infusion test in thediagnosis of reflux gastritis

C B H W l AMERS, F M NAGENGASI, ANI) F 1i M

CORS ITNS (Dept Gas.troceiterologysUniversity Hospital Leiden anid DeptGastroenterology, and Nliclear Medicinie, StRadboud Hospital, Nijmegent, Thle Netther-lands) The diagnosis of alkaline refluxgastritis is based on clinical (post-prandialpain, bilious vomiting, weight loss), endo-scopical and histological findings. Thesccriteria, however, do not allow to reliablyselect patients for surgical treatment. Thealkali provocation test has been reported tobe a simple, safe and rcliablc diagnostictest. To determine the value of this test wehave compared the results of this subjectivetest with the quantitative measurement ofenterogastric reflux of "Tc-HIDA in 16patients (1 I M, 14 postgastrectomy; age35-63 yr) with clinical, endoscopical andhistological findings suggestivc of refluxgastritis. The alkali provocation test wasperformed by intragastric infusion inrandom order of 20 ml salinc, 20) ml 0) 1 MNaOH, and 20 ml () I M HCI separated by10 min periods. The tests were done singleblind and were considered to be positive iftypical complaints were experienced duringinfusion of 0.-1 M NaOH, but not during 0-1M HCI and saline. To determine the repro-ducibility of the test, all infusions weregiven twice. In the "Tc-HIDA test 3 mCiwas injected intravenously and gall bladderemptying was stimulated by oral ingestionof 60 ml corn oil.A positive alkali infusion test was found

in nine of 1) patients with an increasedcnterogastric reflux indcx of more than 100and in two of six patients with a normalindex (p<t-()5), giving an overall agree-ment of 8 l'0O. The tests were fully reproduc-ible in 1(), partly reproducible in four andirreproducible in two patients. Four of thepatients with a positive alkali infusion testhad negative tests when tested afterRoux-en-Y surgery.The alkali infusion test, a simple,

inexpensive and safe test for the diagnosisof alkalline reflux gastritis. shows goodcorrelation with measurement of entero-gaistric reflux using "Tc-HIDA.

Gastric and duodenal subnuclear vacuolatedmucous cells

W TiiOMPSON, I) W DAY, AND N A WRIGiii(Departments of Pathology, Royal Post-graduiate Medical School, HammerstnithHospital and Universitv Department ofPathology, Royal Liverpool Hospital,Liverpool) We report a novel abnormalityaffecting simple mucin secreting cells(SMC) of the glands of pyloric gaistricmucosa Cand of Brunner's glands. Sub-nuclear vacuoliated mucous cells (SVMC)show a distinctive appearance of haema-toxylin and eosin staining. They atrecolumnar cells of similar size to SMC, buthave at central nucleus beneath which thecytoplasm has a uniform 'glassy' eosino-philic aippearance or contains a clear area.Suhnucle,ar vacuolated mucous cells arefound focally lining the lower third ofpyloric gastric glands or in Brunner'sglands, aind may be mistaken for a form ofmetapliasia. Histochemically, the apicalportion ot the cell stains for neutral mucin,but the basal portion stains only weakly andvariably for protein. Electron miscroscopyreveals that the basal portion consists of alarge, single, membratne bound vacuole,variably indented by the nucleus, and prob-ably derived from either endoplasmicreticulum or the Golgi apparatus. Thevacuole contains granular material whichvaries in electron density from cell to cell.The apical portion of the cell contiainsorganelles similar to SMC. These appear-ances are highly suggestive of an abnormalaccumulation of non-glycoconjugatedmucus core protein. Although the cause ofthis is unknown, it wals associated with thehistological changes of chronic gastritis in 1)of our 12 cases.

A prospective study of Campylobacterpyloridis by bacteriology, histology, ureasetest, and serology (ELISA and immunoblot-ting)

A S PENA, N DE VARGiAS, W VAN DUIJN, Pll IiINDTZ, G J A OFFERIiAUS, G, DEN IIARTOG, J

KREUNIN(G, AN[) C B H W LAMERS (Depts ofGastroenterology and Hepatology,Pathology and Microbiology, UniversityHospital Leiden, The Netherlands) Patientsreferred for upper GI endoscopy were

A1408


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Societv of Gastroenterology

evaluated for the presence of C(ampvlo-bacterpyloridis (CP) to determine the valueof different tests in the diagnosis, using thebacteriological test as the gold standard.Twenty six unselected patients were

studied. Five antral biopsy specimens weretaken for culture and bacteriological tests.histology (morphology and Warthy andStarry (WS) staining), and urease test (120-18t) min and 24 h). Serum antibodies againstsonicated CP organisms were assayed by anELISA IgA, IgG. and IgM as well as an IgGimmunoblotting technique (IgG-IB).

Campylobacter pyloridis was found byculture in 5°0% of the patients and fulfilledthe bacteriological criteria for CP. The WSstaining and IgG-IB gave a sensitivity of100)% and a specificity of 85%, ELISA-IgG85% and 91%/0, ELISA-IgA 46%Yo and 85%,histological evidence of gastritis 92% and540%, urease test (120-180 min) 70% and100)/0, urease test (24 h) 92% and 54%° as tosensitivity and specificity, respectively. NoIgM anti-CP were detected in serum.The WS staining technique and IgG-IB

were the best techniques to detect thepresence of CP, followed by ELISA-IgG.The urease test was less satisfactory. Noneof the patients but one with normal histo-logy had a positive culture. Three patientswith chronic antrum gastritis had a negativeculture and WS staining but positive anti-bodies, thus suggesting, a possible aetio-logical role of CP in chronic antrum gastritisin addition to the well established CP role inacute antrum gastritis.

Gastric secretion of antibiotics used forCampylobacter pyloridis

J A iiOL,INGSWOR1Tii, J (GOiLDIE, C F SIllEi'll', Y

I,, H RICIIARDSON, AND R H HUNT (McMasterUniversity Medical Centre, Hamilton,Ontario, Canada) Erythromycin has a lowrate of eradication of Campylobacterpyloridis (150%) while metronidazole andampicillin csters are highly effective (85%).We have found C pyloridis difficult toeradicate from the gastric glands and postu-late that while topically active agents mayclear the organism initially early relapsemay be associated with persistance in gastricpits. Gastric secretion of antibiotics shouldresult in high drug concentrations in thegastric glands. Gastric juice levels ofmetronidazole and ampicillin weremeasured in four subjects in samples takenone hour before and two hours after thedrug were given intravenously. Ampicillinwas measured by bioassay and metronida-zole by HPLC. No ampicillin was detected

in any samples despite a mean serum levelof 4 95 mcg/mi (sd 0( 12). Peaik metronida-zole levels in gaistric juice werc 61 [tg/ml (sd8) at 3t) mins aind 30 tgr/ml (sd 0(23) inserum at one hour.These findings suggest that gastric secre-

tion of antibiotics does not explain theirvarying efficacy in the eradication of Cpyloridis. Other factors, such as luminaleffects of oral atntibiotics on gastric mucusand mucosa. and tissue levels may be moreimportant.

In vitro mucus glycoprotein synthesis andsecretion by gastric mucosa colonised withcampylobacter pyloridis

J E CRABTRIE, B J RAlU'IBONI-., J WYA'T, R V

IIEATLEY, AND M S l.OSOWSKY (Dep)t of)Medicine and Dept of Pathologs, St James 'sUniversity Hospital, Leeds) There is astrong correlation between Camp(yobhacterpyloridis colonisation and gastritis. Theassociation between surface C pyloridiswith depletion of neutral cytoplasmicmucins in the gastric epithelial cells raisesthe possibility that C pyloridis may bedirectly or indirectly affecting the protectivemucus layer in the stomach. To study thecapacity of normal and gastritic mucosa tosynthesise and secrete mucus glycoproteins,24 hour in vitro cultures of antral and bodybiopsies were undertaken. The incorpora-tion of "H glucosamine into tissue anidsecreted glycoproteins was measured(DPM x I 0/mg protein).C pyloridis colonised. antral biopsies

secreted significantly higher quantities oflabelled glycoproteins than normal antraltissue. Mean values respectively were106±14-2 (n=6) and 58-7±98- (n=7)(p<0(05). Conversely, tissue associatedlabelled glycoproteins were higher innormal antral biopsies (198± 19-2) thangastritic biopsies (159+ 15 5). Totalincorporation of glucosamine into tissueand secreted glycoproteins was significantlygreater (p<0.0l) in C pvloridis infectedbody biopsies (258± 19 2; n=7) than normaltissue (146±10(3; n=8). The increase inglycoprotein synthesis correlated with anincrease in epithelial cell numbers (r=().66,p<005). The ratios of tissue to secretedglycoproteins for normal and gastritic bodytissue were 2-8:1 and 17:1 respectively.These results show that in non-

autoimmune gastritis where C pyloridisis present significant alterations to gastricmucus production occur. Whether thealterations in mucus production are a directaffect of C pyloridis remains to bedetermined.

Efficacy of different dosage regimes induodenal ulcer healing and eradication ofCampylobacter pylori

J (G (C(GiILAN. I) GIll(iAN, If iiUMPiiRYS. 1)

MCKI-NNA. F SWEiENEY, C K-.ANI.. ANi) (O MoRAIN (Dept of Micro (111(d Histopath, SIJames( Hospital andit Trinity (ollege, Dept of'Gastro, MetathiA dtelaide Hospitals, Dlblin,Irelandti) C(ampvylobacter pYlori (CP) statusof piatients with healed duodenali ulcers hiasrecently been shown to be highly predictiveot the likelihood of relaipse. Colloidallbismuth subcitrate (CBS) is active againstCP in vivo. The atim of this prospectivecontrolled triill was to assess the efficaicy ofCBS 120 mg qds v 240 mg bd in ulcer healingand CP eradication.

Sixty consecutive piatients with endo-scopic duodenal ulcers (DU) randomised(30 bd 30 qds CBS). Repeait endoscopy wasperformed at four weeks and again at eightweeks in patients who haid not heilled at fourweeks. At each endoscopy two antral biop-sies were taken from <2 cm from thepylorus and assessed histologically andmicrobiologically for evidence of CP.

At entry 90%) of both groups werepositive for CP, at four weeks 62%' of bdand 33%, of qds patients remained CPpositive (p<0-05), at eight weeks 50%, of bdpatients and 16% of qds patients remainedCP positive (p<0-02).At the end of the study period there was

no significant difference in healing ratesbetween the groups (65%, of bd and 75('o ofqds having healed).Ods dosage reduces CP infection more

effectively and should he associated with alower DU relapse rate. Different mecha-nisms of action must operate in CP eradica-tion and DU healing with CBS.

Campylobacter pyloridis in tropical Africa

JUDITlI I WYATT, J S DE CAESTFCKER, B J

RAItIBONE, AND R V HIEATiLEY (Depts Patho-logy and Medicine, St James's UniversityHospital, Leeds and Dept Medicinte,University of Science and Technology,Kumasi, Ghana) Virtually all Campylo-hacter pyloridis (CP) studies to date haveconcentrated on western style populations,yet in the tropics duodenal ulcers arereported to be particularly common andbenign gastric ulcers rare. We studied thepresence of CP, its relation to mucosalinflammation and patients serologicalresponses in 39 patients being endoscopedin Kumasi, Ghana, West Africa. Histologywals performed on antral, body, and duo-

A 1 409(


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


denal biopsies including a modified Giemsastain for identifying CP. Serum antibodytitres to CP were studied by ELISA using asoluble antigen preparation. The patientsmean age was 41 years (range 12-71) (male:female 28:1 1). The mean duration ofabdominal symptoms was four years. Thirtyeight of 39 patients had antral gastritis, and23 of these (59%) had duodenal ulcer (14)or duodenitis (nine). The 38 gastriticpatients were all CP +ve, and the singlenormal subject -ve. The prevalence of CPcolonisation in this African population isthe highest reported in any study. Thestrong association with CP gastritis andduodenal disease seen in the Ghana patientsis similar to that observed in the UK.

The effect of vagotomy on Campylobacterpylori

BELINDA J JOHNSTON, P I REED, AND M H Al.I(Departments of Gastroenterology andHistopathology, Wexham Park Hospital,Slough, Berkshire) It is recognised that thegastritis seen in duodenal ulcer (DU)patients is associated with the presence ofantral Campylobacter pylori (CP). It hasbeen suggested that after gastric surgery thegastritis type will alter and may no longer beassociated with CP presence.To establish the existence of such a differ-

ence, gastric antral biopsies were takenfrom 61 patients who underwent vagotomyfor DU, 1-20 years previously, and from 75untreated patients with active DU. Of thosewho had surgery 27 had highly selectivevagotomy (HSV), 26 vagotomy and pyloro-plasty (VP) and eight truncal vagotomy andgastroenterostomy (TVGE). The biopsieswere assessed histologically and examinedfor the presence of CP. Gastric juice pH wasmeasured in the vagotomy patients.

Campylobacter pylori were present in 72(96%) of the DU patients all of whom hadhistological gastritis. Forty four (71%) ofthe vagotomy patients were CP positive and41 of these had histological gastritis (HSV67%, VP 65%, and TVGE 75/), whichtended to be more pronounced in theTVGE group. The severity of the gastritiswas unrelated to the intragastric pH or timesince surgery. Although the proportion ofnormal biopsies was higher in patients withvagotomy than with active DU, the gastritisin the former group was always associatedwith the presence of CP. These data wouldindicate that after vagotomy, the associa-tion between CP presence and gastritis issimilar to that found in the unoperated DUstomach.

Non-invasive measurement of gastric empty-ing using applied potential tomography(APT)

I) F FVANS, J W WRl(IlG, (G LAMONT, AND J D

iIARDCASTiLF (Department of Surgerv,University Hospital, Nottingham) Appliedpotential tomography is a new non-invasivetechnique which measures changes in tissueresistivity via surface electrodes. Gastricemptying of meals can be assessed using thismethod without the need for radioisotopesor expensive nuclear imaging equipment.

Five normal subjects were imaged on twooccasions after ingestion of 500 ml Oxodrink and again after 500 ml of Quaker Oatsporridge to assess liquid and solid emptying.Resistive images were collected at oneminute intervals for 60 and 180 minutesrespectively. The tests were repeated after40t) mg of cimetidine as gastric acid secre-tion is thought to influence imaging usingAPT.The time taken to empty 5000 of the test

meals was expressed as ti,. In the controlstudies ti, for the liquid meal was a medianof 35 min (range 20-40 min) compared withti, of 19 min (range 18-31 min) with cimeti-dine (p=)-028). For the solid meal ti, con-trol was a median of 102 min (range 80-117min) compared with tin 68 min (range 43-94min) with cimetidine (p=0016).

Applied potential tomography is a uscfultechnique to evaluate gastric emptying andis particularly suitable when multipleinvestigations are required. Cimetidinedoes modify the emptying pattern and thisshould be considered for future studies.

Enterochromaffin like (ECL) cell popu-lations are dependent on age and sex

A E BISHOP, D M GRFENiG RINDI, F I l.EE, P J

ISAACS, M J DALY, J DOMIN, S R BliOOM, ANI) J M

POLAK (Depts of Histochemistrv andMedicine, RPMS, Hammersmith Hospital,London, Dept of Pathology, Leeds GeneralInfirmary, Leeds, Dept of Gastroenter-ology, Victoria Hospital, Blackpool, andAstra Pharmaceuticals, Edinburgh) Toexamine variations in ECL cells with ageand sex, a two part, retrospective andprospective, study was made. Retrospec-tively, surgical and endoscopic biopsies,previously reported as normal althoughsome had low grade gastritis, were takenfrom the files. Four groups of subjects werestudied; men 45 (n=5) and 55 years (n= 13)and similar female groups (n=5) andn= 15). Prospectively, endoscopic biopsieswere taken at standard sites from stomachs

of selected subjects without gastritis (men45, n=9, 55, nr=l10 women 45, n=-l, 55n=8) and plasmat gastrin concentrationsmeasured. In the retrospective samples, theonly significant change was a rise in ECLcells of older females (13-6±1.9, mean+SEM, cells/visual field) compared with theyoung group ( t)-6+±0.8, p<0)-0)3). The pros-pectively biopsied, however, selected sub-jects of both sexes showed reduced ECLcells with age (mcales 45, 50)+ 106, 55,33+3-6. p<)-05:, females 45, 44±6o5, 55,31+±44, p<0)05). Gastrin concentrationsdid not change within sexes but were higherin older females (229±+7-5 pmol/l) com-pared with older men (63+±0-6, p<0O05).The difference in the two sets of results mayreflect the high incidence of gastritisobserved in non-selected older women.This in turn may relate to plasma gastrinconcentrations which were relatively higheven in the selected group of older females.These variations should be taken intoaccount when assessing pathologicalchanges in ECL cells.

Duodenal acidification and gastrin responsesafter feeding: a new look at the patho-physiology

C A ERIKSEN, K D BUCHIANAN, AND A CUSCIIFERI(Department of Surgery, Ninewells Hospitaland Medical School, Duniidee and Dept ofMedicine, Queen 's University (f' Belfast,Belfast) Controversy exists regardingthe duodenal acidification and gastrinresponses to sham feeding and a meal inulcer disease. We measured plasma gastrinand duodenal bulb pH before, during, andafter modified sham feeding (MSF) and asolid meal in 16 duodenal ulcer patients(DU) and 12 volunteers (VOL). Gastrinlevels were significantly higher in DU(fasting: DU 42 5 ng/l, VOL 22-5 ng/l,p<0)-001 ; meal peak: 130)(, 60-0, p<0(02).Two distinct DU groups emerged: 'hyper-gastrinaemic' (HrG) showed exaggeratedgastrin responses; 'normogastrinaemic(N"G) showed levels similar to controls.Cephalic stimulation produced significantlygreater responses only in H'G (HiG fasting67-7 ng/l, MSF peak 13t)() ng/l, p<t)005).

After the meal, both patient groupsexperienced significant delays in both, onsetof duodenal acidity (H'G: 54-0 min, NG:89-5 min, VOL: 16.0 min, p<0(02), andtime to peak acid response. HrG showedprolonged duodenal acid exposure andreturned to pre-meal pH levels significantlyslower (h'G: 78t) min, N"G: 28X0 min,p<0.005).

A 1410


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


These results show a hypergastrinaemic'subset of DU patients. exhibiting markedcephalic responses, abnormal duodenalcidification and a defective switch offmechanism of acid secretion.

Evidence for a gastro-cholecystic reflex forgall bladder emptying in man

S E'LLENBOGEN, J S GRIME, J CAIAM, C R

MACKIE, S A JENKINS, AND J N BAXTI.R(Department of Surgerv, University ofLiverpool, Department Of NuclearMedicinie, Royal Liverpool Hospital,Department of Gastroenterology Roval PGMed School, Lonidoni) A gastric phase ofgall bladder emptying (GBE) has beendemonstrated in the dog. In this study "Tc'-EHIDA cholescintigraphy was used to con-firm its presence and mechanism ot action inman. Healthy volunteers were studied;group 1, no gastric distension (n= 18); group11, gastric distension by balloon (n =9);group 111, atropine given before gastricdistension (n=7); group IV, meal ingestion(n= 14). Gastric distension was also per-formed in patients with truncal vagotomygroup V (n=9). Blood samples were takenfor cholecystokinin bioassay. In group 1, theprobability of spontaneous GBE in anyminute period was 0(0053. Significant GBEoccurred in six of nine (66%0) group 11(p<0.00(l, group II v group 1, Poisson Test),none of seven (0%) group III (p=(-0144,group III v group II, Fisher's Exact Test), 14of 14 (1(()%) group IV, four of nine (44,o)group V (p=0-6372, group V v group 11,Fisher's Exact Test). From the start ofgastric distension, onset of GBE was8X2±1-9 minutes group II, 11 9±2+6minutes group IV, 4-8+2-1 minutes groupVI. From the start of GBE, 2t) minute gallbladder ejection fraction was 41'0+±69°(Ygroup 11, 29 2±5+3% group IV, 32 2+5 8X°ogroup V. Plasma cholecystokinin did notrise significantly from basal followinggastric distension.These results suggest that a gastric phase

of GBE exists in man, which persists aftertruncal vagotomy but is inhibited by atro-pine and independent of cholecystokinin.

The 'acid antrum' as a possible cause ofrecurrent ulceration after highly selectivevagotony (HSV): a histological study

K S NAIK, M l AGOPOUl.OS, J N PRIMROSE, AND I)JOHNSTON (University Departments ofSurgerv' and Aniatomy, The University,Leeds and Leeds General Infirmary, Leeds)

The aim of this study was to standardise thedefinition of the antrum in terms of themacrocopic antrum - corpus boundarydefined by the nerve of Latarjet (NL-ACB)and microscopic ACB according to bothparietal cells (PC) and gastrin cells (GC),and hence determine if HSV providesreliable denervation of the distal PC mass.

Serial sections obtained from longitu-dinal strips of 43 human 'normial' post-mortem stomachs were stained for PC (43stomachs) and GC (in 20 of the 43). The PC-ACB was distal to NL by more than I cm in26 stomachs (61'%). In seven stonmachs(16'%) PC extended to the pylorus ('acidantra'). Of the 2t) stomachs stained for GCand PC. PC extended to the pylorus in four(20o). In these, GC were sparse and theantrum as defined by GC was smaller(p<0)-01). The GC-ACB correlated muchbetter with NL-ACB than did PC-ACB.These findings suggest that current tech-

niques of HSV often do not provide reliablePC denervation, particularly in patientswith 'acid antra'. T'hese patients may beidentified by preoperative histologicalexamination of antral biopsies or peropera-tive Grassi test.

Effect of indomethacin on human gastro-duodenal 'mucus-bicarbonate' barrier

C J SHJORROCK AND W D W REES (HopeHospital, SalJord) We have studied theeffect of indomethacin on the 'protective'mucus-bicarbonate barrier overlyinggastroduodenal mucosa in man. The pHgradient across fundic, antral and duodenalmucus gel was measured by a pH micro-electrode inserted down an Olympus 010gastroscope. Mucosal integrity was gradedendoscopically from () (normal) to 4 (severedamage) according to Lanza. In six healthysubjects (aged 22 to 35 years) with normalmucosa a pH gradient was measurable infundus (maximum intramucus pH=6-4±0-16), antrum (5-7±()-25) and proximalduodenum (66±0+1). Twenty four hoursafter starting indomethacin (50 mg tds) allsubjects had mucosal damage (score:gastric=2-0±)0-36; duodenum t)-66±t)-33).At seven days of aidministration, damagepersisted (gastric= 1X83+±)56; duodenal0-66±0-66) and two subjects developeddiscrete ulceration (one gastric and oneduodenal). In the three subjects studied at28 days of administration, all mucosaldamage had resolved. Intramucus pHremained unaltered at 24 hours and dayseven, with only fundic mucus pH beingslightly increased at day 28 (6-9±0(07,

p<)-t)5, n=3). Antral luminal pH wassignificatntly raised alt 24 hours (3-35±t)-59compared with l-81±t)15. p<)')2, n=6)returning to normal by day 28.

Indomethiacin damages mucosa withoutaltering intramucus pH aind there is evid-ence of mucosal adaptation with continuedadministration.

Acidification of the perfused human stomachduring muscarinic-M l-receptor blockadestimulates gastric prostaglandin (PG) E2output

A MiERTI-NIEIiSEN, 1 K MUNCK, K BUKIIAVI,AND J RASK-MADSI N (Departmlnens oJMedical Gastroenterology, Bispebjerg andHerlevy Hospitals, University oJ' Deintark)Pirenzepine dose-dependently increasesduodenal HCO 3 secretion, in addition toinhibiting gastric H' secretion. Further-more, duodenal aicidification stimulatesHCO A and PGE, output. To study theinfluence of gastric luminal acidification onmuscairinic regulation of gastric PG release'steady state' perfusions of the stomnachwere carried out in eight healthy volunteersduring 'shamteeding' before and after oraladministration of pirenzepine (5t) mg bid;four days). Luminal acidification (30) mMHCI) slightly decreased gastric peak acidoutput (PAO) (1-+(0-5 v control 2-6±t)3mmol/15 min; mean+SEM; p<t)'05), butcaused no change in luminal PGE. release(10(5±3- 1 v control 8X2±+15 ng/15 min).Pirenzepine reduced PAO (0(9±+()2;p<()'()5) and a small rise in PGE. output(13-9±3-2; p<0-0)5) was observed. Luminalacidification further decreased PAO(0(7±0)3; p<t)-()5) and markedly increasedPGE, release (21-9±6-6; p<0)'()5). Thisresponse was abolished (H't 1.1±0_4;PGE.: 2-9±()-8; p<()-05) by indomethacin(1()0 mg iv). In conclusion, gastric mucosalPGE. output was significantly increased byluminal acidification during muscarinic(MI) receptor blockade. The underlyingmechanism may be important for the under-standing of 'cytoprotection' and couldexplain why pirenzepine heals ulcers betterthan predicted from its acid inhibitorycffect.

Polyunsaturated fatty acid ingestion andpeptic ulcer - evidence for a direct relation-ship

H W (,RANI, R W KELLY, K R PAIMIR, AND) J

MISIEWIC Z (Gastrointestinal Unit, WesternGenieral Hospital and MRC Department of

A 141 1


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/


Reproductive Biology, Edinbt,rgh) Theincidence and virulence of peptic ulcerdisease has fallen in the Western hemis-phere, as consumption of polyunsaturatedfatty acids has increased. We havcexamined the effects of dietary poly-unsaturated fat on gastric atcid secretion,prostaglandin output and serum gastrinconcentration. Nine healthy volunteerstook 1-5 g or 3 g of linoleic acid (LA) for14-21 days. Linoleic acid did not affectgastroduodenoscopic appearances, gastro-duodenal histology, routine h(aemratologynor serum and urinary biochemistry. Foursubjects developed mild diarrhoea. Meanbasal gastric acid output was 8-3±1-5(SEM) mmol/h before LA and 7-2±1-3(SEM) mmol/h after LA (p>0(05). Meanpentagastrin stimulated gastric acid outputdecreased from 35-2±3-5 to 30-1 ±2-8mmol/h (p<0-05). There was a correspond-ing increase in mean serum gastrin concen-tration from 19 22±3-7 to 30-89±3-8 pg/mI(p<0-0l). The mean output of immuno-reactive PGE in gastric juice increased from498±110 (SEM) ng/h to 123(0±475 ng/h(p<0-05). The mean output of the main

metabolite ot PGE - 13,14-dihydro 15-ketoPGE, in gastric juice increased from192±19 (SEM) to 1335 (+708) ng/h afterLA (p<0(01). These data suggest anexplanation for the observed relationshipbetween dietary polyunsaturated fatty acidintake and peptic ulcer disease; poly-unsaturated fatty acid ingestion modifiesgastric prostaglandin metabolism withsecondary changes in acid secretion andgastrin output.

A prospective, controlled trial of the garrengastric bubble (GB) with or without 600calorie diet (D) on weight loss and gastricemptying (GE)

N (GEMAYIE, I.AINI., 11 COHIEN, N ARNSIEIN,AND M BOOSAI IS (INTRoDuCrE BY JVALEN/ZULLA) (USC Schiool of Medicine,Los Atigeles, C(A, USA) Controversy existsregarding the efficaicy and physiologiceffects of the GB. Sixteen obese women(age 3(0-58, BMI 32-52 kg/m') have been

randomised into one of four groups (1) NoGB, No D (n-4); (2) GB, No D (n=4); (3)No GB, D (n=3); (4) GB, D (n=5) aindhavc completed a three month trcatmentperiod. All patients had endoscopy withreal or sham GB insertion. Three monthweight loss (kg): (1) -)-9±+13, (2)-0)-9± 13, (3) -5-9±3+1, (4) -11-3± 1*3;p<0-05 for variation among the meanweight losses and for I v 4 and 2 l} 4. LiquidGE wais significantly more rapid two wceksafter GB insertion as compared to pre-GB:ti,: 11±2 v 19±3 min; time to duodenalvisualisation: 3± 1 v 7± 1 min; {MO residual at3(0 min: 17±2 l 32±5%/O. A significantcorrelation wis seen between weight lossand the ti, for liquids after GB insertion(r=(182, p-(0007). Although aIn erraticsolid GE curve developed aftcr GB inser-tion, quantitative pairiameters of solid GEwere not significantly difterent.

Interval results of this ongoing trialsuggest ( I ) dietary instruction is the keyfactor in this weight loss prograrm, (2) GBwithout dietary therapy is ineftective, and(3) liquid GE is significantly haistened aifterGB insertion.

A 1412


http://gut.bmj.com

/G


ctober 1987. Dow

nloaded from

http://gut.bmj.com/

The British Society of Gastroenterology - Gut

Documents

Transcript of The British Society of Gastroenterology - Gut