PancreaticSociety ofGreat Britain andIreland · patients undergoing surgery for chronic pancreatitis. Thehistology has beenevalu-ated by an experienced liver histopatholo-gist and

Gut, 1988, 29, A265-A272

Pancreatic Society of Great Britain and IrelandThe 12th Annual Meeting of this Society was chaired by Mr Michael Knight MD, FRCS, at St George'sHospital on 6 November, 1987. The main Guest Speaker was Professor J H C Ranson from New York. LordSmith of Marlow presented the first Rodney Smith Prize for the best paper to Mr John Neoptolemos for thefirst class work on the role of ERCP in gall stone acute pancreatitis; and the second Travelling Fellowship ofthe Society was awarded to MrM Larvin of Leeds. Additional guest speakers were Dr Peter Trott (pancreaticcytology) and Mr Christopher Russell (total pancreatectomy). Abstracts of the main papers are printedbelow.

Bile-salt/lysolecithin induced model of acutenecrotising haemorrhagic pancreatitis

A L BLOWERt, T V TAYLORt, AND R F T

MCMAHON* (tDepartmentofSurgical Gastro-enterology, Manchester Royal Infirmary,Manchester, and *Department of Patho-logy, University of Manchester MedicalSchool, Manchester) A model of acutehaemorrhagic pancreatitis (AHP) involvingcannulation and infusion of substances intothe pancreatic duct in a retrograde fashionhas been devised which obviates the use oflarge, unphysiological volumes of infusateat high pressures, and avoids ductal rupture.

In the male Sprague-Dawley rat (weight350-400 g), AHP was induced by transduo-denal cannulation of the bile pancreaticduct (BPD) and slow infusion (over threeminutes) of 50 tl of a mixture of 7-5%glycodeoxycholic acid and 1% lysolecithin.The proximal BPD was occluded during theinfusion and the distal BPD ligated afterremoval of the cannula. The ductal pressureremained within physiological limits duringthe infusion at 11-3±0-9 cm of water.

Macroscopic pancreatic haemorrhagesoccurred within minutes, the developmentof blood stained ascites by 15 minutes, and aconcurrent rapid rise in serum amylase andlipase. There was a fall in systemic bloodpressure within one hour. Histologically,marked pancreatic oedema with polymorphmargination of capillaries occured within 15minutes. At 30 minutes there was obviousdegranulation of the acini, but the glandarchitecture remained intact throughout.At one hour extensive and irreversiblechanges of pancreatic necrosis werepresent.When the same volume of Indian ink was

infused no leakage from the duct systemoccured suggesting that ductal rupture playsno part in inducting pancreatitis in thismodel.The mortality in 20 animals at 24 hours

was 60% and at 72 hours 90%. This is the

only simple, easily reproducible model ofAHP, which uses a physiological volume ofinfusate, injected at physiological pressure.

Prophospholipase A and trypsinogen activa-tion are not responsible for the initiation ofhyperstimulation pancreatitis in mice

K SHORROCK, P R HURLEY, AND J HERMON-

TAYLOR (Department ofSurgery, St George'sHospital Medical School, London)Oedematous pancreatitis induced bysecretagogue hyperstimulation is widelyused as a model of non-lethal oedematouspancreatitis. Despite extensive investiga-tion the initiating mechanisms of thisdisease are unknown. This study was under-taken to determine the role of inappropriateenzyme activation in the initiation of tissuedamage.

Pancreatitis was induced in male CBAmice, weighing 20 g, by intermittent hourlysubcutaneous administration of CCK8 at adose of 100 ,tg/kg/h for up to six hours.Animals were killed at intervals for eighthours. Blood was collected for serumamylase assay. The pancreas was dividedinto two equal portions. One half wassubjected to radioimmunoassay of the acti-vation peptide of trypsinogen (TAP). Theremaining portion was homogenised andactive phospholipase A (PLA) assayedusing radiolabelled dipalmitoyl lecitin as asubstrate.Trypsinogen was not detected at any time

during the period of study. Similarly, thelevel of PLA in the pancreatic extract didnot rise above control levels. PLA activitywas not affected by adding aprotinin to theincubation medium.

It is concluded that neither trypsinogennor prophospholipase activation areresponsible for the early pancreatic damageobserved in hyperstimulation pancreatitis.

A265

Gastric function after conventional andpylorus preserving pancreatoduodenectomy

M J COOPER, E RHYS-DAVIS, AND R C NWILLIAMSON (University Departments ofSurgery and Radiodiagnosis, Bristol RoyalInfirmary, Bristol) Preservation of thepylorus during pancreatoduodenectomy(PP) was reintroduced in 1978 to avoid thepostgastrectomy syndromes that accom-pany conventional Whipple resection(CW). Since 1980 we have performed PP on26 patients, of whom 17 (8M:9F, medianage 48 yrs) have subsequently undergoneisotopic assessment of gastric function; theresults have been compared with 11 CWpatients (7M:4F, median age 55 yrs) fromthe same period. To assess gastric emptying(GE), patients were given a meal of twoscrambled eggs radiolabelled with 12 MBqtechnetium sulphur colloid (solid phase;normal half-time <90 min) and 300 ml milkcontaining 12 MBq of Indium111DTPA(liquid phase normal half-time <45 min).Counting was performed by an IGE widefield gamma camera for three hours or untila half time was achieved. After PP GE wasnormal in five (29%), moderately delayed(up to twice normal) in 10 and grosslydelayed (>200 min for solids) in two; onlythe latter two patients were symptomaticand both required revision surgery. OneCW patient emptied precipitately (<4 min)and one was normal, six showed moderateand three gross delay. Competence ofthe preserved pylorus was assessed in 12patients following IV injection of 80 MBqtechnetium-labelled BIDA; milk and CCKwere given after 30 min and imaging con-tinued for a further 60 min. Nine patientshad no or only minimal reflux, but three hadreflux into gastric body or fundus. All 8 CWpatients tested showed reflux, four into theremaining fundus. GE is more likely to benormal after PP and the intact pylorusreduces bile reflux. For most patients gastricstasis is not a clinical problem.

on August 21, 2020 by guest. P

rotected by copyright.http://gut.bm

j.com/

Gut: first published as 10.1136/gut.29.2.A

265 on 1 February 1988. D

ownloaded from

http://gut.bmj.com/

A266 Pancreatic Society of Great Britain and Ireland

Guided percutaneous pancreatic biopsy

C J MITCHELL, A M JACKSON, D WAI, ANDJ MACFIE (Scarborough Hospital, NorthYorkshire) A major problem in the diagno-sis and assessment of pancreatic disease hasbeen to obtain histology without resort tooperation. Fine needle aspiration biopsyproviding cytological material for theexpertise in interpretation is not widelyavailable. Initial experience using largergauge needles has been encouraging. Abiopsy suitable for histological examinationis obtained obviating the need for specialistinterpretation. We report our initial experi-ence with this technique.

Ultrasound guided percutaneous pancre-atic biopsy was performed using an 18 swgTru-cut needle activated by a fast actingfiring mechanism (Radiplast, Sweden)which has several practical advantagesparticularly for accurately positioning theneedle. Adequate samples were obtained in21 of 23 attempts. Chronic pancreatitis wasdemonstrated with no evidence of neoplasiain nine patients in whom a pancreatic masswas seen at ultrasound (seven) or in whomthe diagnosis was in doubt (two); in afurther patient with CP the biopsy showedan islet cell tumour. The clinical and ultra-sound diagnosis of pancreatic carcinoma(nine patients) was confirmed in sevenpatients and lymphoma and an eroding gas-tric neoplasm in the other two patients.Biopsy of the dorsal anlage in two patientswith pancreas divisum showed normaltissue and pancreatitis respectively. Onepatient experienced considerable pain for48 hours after the procedure but no othercomplications were encountered.

If this technique proves to be as safe andaccurate as initially suggested, it providesan important new diagnostic tool in thestudy of pancreatic disease. The suitabilityof the biopsies for histological examinationmeans that the technique can be used out-side specialist centres.

Reference1 Hall-Craggs, MA, Lees WR.Gut 1987; 28: 233-5.

Incidence of hepatobiliary complications inpatients undergoing surgery for chronicpancreatitis

C WILSON, R SCHLINKERT, C D AULD, A H

HASAN, C W IMRIE, R M N MCSWEEN, AND D CCARTER (Department of Surgery, RoyalInfirmary and Department of Pathology,Western Infirmary, Glasgow) Common bileduct stenosis (CBDS) complicating chronicpancreatitis may give rise to obstructive

jaundice and rarely secondary biliary cirr-hosis. CBDS may also be manifested byasymptomatic alkaline phosphatase (ALP)rise,' the incidence and significance of whichis unknown. We have undertaken operativeneedle biopsy of the liver in 39 consecutivepatients undergoing surgery for chronicpancreatitis. The histology has been evalu-ated by an experienced liver histopatholo-gist and the findings correlated with pre-operative radiology and biochemistry.

Features of extrahepatic biliary obstruc-tion were found in 12 patients. Threepatients had parenchymal liver disease(alcohol related in one) and two others hadfeatures suggestive of previous alcoholinduced injury.

All nine patients with ALP >2x normaland successful biliary radiology showedCBDS but only six (67%) had features ofextrahepatic biliary obstruction on liverbiopsy. Five (83%) of the patients withjaundice, eight (50%) with CBDS and eight(67%) with ALP >2x normal had featuresof extrahepatic biliary obstruction on liverbiopsy. ALP >2x normal suggests thepresence of significant CBDS but neitherindex provides a definite indication forbiliary by-pass, and preoperative liverbiopsy may be valuable in this context.

Reference1 Snape WJ, Long WB, Trotman BW, Marin GA, CzajaAJ. Marked alkaline phosphatase elevation with partialbile duct obstruction due to calcific pancreatitis. Gastro-enterology 1976; 70: 70-3.

Results of a prospective randomised trial ofERCP and endoscopic sphincterotomy (ES)in acute pancreatitis caused by gall stones

J P NEOPTOLEMOS, D L CARR-LOCKE, N

LONDON, I BAILEY, D JAMES, AND D P FOSSARD

(Departments of Surgery and Gastro-enterology, The Leicester Royal Infirmary,Leicester) Between 30% and 60% of

patients who die from gall stone associatedacute pancreatitis have stones present in thecommon bile duct (CBD). ERCP, for diag-nosis, and ES, for stone extraction, have notbeen properly assessed in the urgent situa-tion because of a feat of procedural compli-cations. We report the results of a prospec-tive randomised trial to assess the value ofthese techniques.Over a five year period, from 223 con-

secutive patients admitted with acute pan-creatitis of all causes, 131 patients withsuspected pancreatitis were randomised toreceive either conventional treatment or toundergo urgent (<72 hours) ERCP; ES andstone extraction was undertaken only ifCBD stones were found at the time ofERCP. The diagnosis was based on a serumamylase >1000 IU/l in the presence of acompatible clinical picture. Gall stoneswere diagnosed by urgent ES or bio-chemical prediction. Patients were prospec-tively stratified using (modified) Glasgowcriteria of severity. Ten patients wereexcluded because an alternative aetiologywas subsequently found. The results in theremaining 121 patients are shown in theTable.The complication rate was significantly

greater in the conventionally treated groupconsidering all cases (x2=9 35, df=1,p=0-002). There were significantly fewercomplications in patients with confirmedstones and predicted severe attacks who hadERCP/ES (x2=4-93, df=1, p=0-026), butnot in those with predicted mild attacks. Ofthose undergoing ERCP, CBD stones weremore frequent in those with predictedsevere attacks (63%) compared to thosewith predicted mild attacks (25%; X2=481,df=1, p=0-029).These results show that urgent ERCP+

ES is a safe and rational approach to themanagement of patients with severegallstone-associated pancreatitis resultingin a significant reduction in the overallcomplication rate.

Complications Patients withcomplications

Group n Pseudocysts Systemic Deaths (n)

Mild - conventional(*)GS -confirmed 29 4 0 0 4(14%)

GS - not confirmed 5 0 0 O0Mild- ERCP/ES

GS-confirmed 28 3 1 0 4(14%)GS - not confirmed 6 0 0 0 0

Severe - conventionalGS-confirmed 24 6 9 3 13GS- not confirmed 4 2 3 2 4

Severe- ERCP/ESGS - confirmed 22 3 1 0 4(18°/GS - not confirmed 3 0 2 1 2(18%)

*GS=gall stones.



j.com/



ownloaded from

http://gut.bmj.com/

Pancreatic Society of Great Britain and Ireland A267

Cytosolic retinoic binding protein in humanpancreas

J K JUTLEY, J KELLEHER, T G BRENNAN*, C J

MITCHELL, AND M E DENYER (*Departmentsof Medicine and Surgery, St James'sUniversity Hospital, Leeds) Vitamin A andits analogues are important regulators ofepithelial cell differentiation. Retinoidshave been shown to inhibit growth ofvarious cell lines, in cell cultures, by directinhibition of cell proliferation. Themolecular mechanism of action is poorlyunderstood. It is thought to mediate itsactions by specific intracellular bindingproteins. The aim of this study was to assesscytosolic retinoic acid-binding protein(cRABP) in the human pancreas.

Tissue from three normal, seven chronicpancreatitis and three patients withpancreatic carcinoma was collected on ice,immediately frozen in liquid nitrogen andstored at -70°C. The tissue was homo-genised, centrifuged at 100000 g for onehour and the cytosolic fraction incubatedwith [11, 12-3H] transretinoic acid in thepresence or absence of 100-fold excessunlabelled ligand. The incubate was thensubjected to either saturation analysis orsucrose density gradient centrifugation.

Using sucrose density gradient centrifu-gation, cRABP values varied from 0-7 to 21pmol/mg in chronic pancreatitis and pan-creatic carcinoma tissue. There was nodetectable cRABP in normal tissue. Therewas a highly significant difference betweenchronic pancreatitis and carcinoma of thepancreas.

Scatchard analysis revealed a dissociationconstant (Kd) of 21 5±5 7 (SE) nM andmaximum binding capacity of 5 2±2-6 (SE)pmol/mg protein.A single class of high affinity binding

protein was present in the pancreas.cRABP was present in all the diseased, butnot in the normal, pancreas. The highestlevel of receptor were found in carcinoma ofthe pancreas.

Effect of intestinal operations on thepancreas

I STEWART, B FLAKS, M J COOPER, P W DAVIES,AND R C N WILLIAMSON (Department ofSurgery, RPMS, Hammersmith Hospital,London) Several experimental models haveshown a role for pancreatic exocrine growthin the development of pancreatic adeno-carcinoma. Understanding this cocarcino-genic role requires a knowledge of thecharacteristics of growth mechanisms.

We have used two established rat modelsto show differences in growth and histology.Male Wistar rats underwent pancreatico-biliary diversion (PBD, n=15) and exten-sive proximal small bowel resection (PSBR,n= 15). Control rats (n= 17) had laparotomywith triple transection of the intestine andreanastomosis. At six months, the animalswere killed, pancreatic wet weight measuredand histological studies performed.Mean pancreatic weight (mg panc/100 g

bw) was 2.19 for controls, 4.45 for PSBR(p<0.001), and 5-24 for PBD (p<0.001).The increase observed in the PBD groupwas greater than that in the SBR group(p<0005). Acidophilic acinar cell foci(putative premalignant lesions), were notedin the pancreas of all rats in the PBD group,whereas they were virtually absent in theSBR and control groups.Both operations therefore stimulate pan-

creatic growth, but with different intensitiesand histological patterns, indicating theexistence of more than one endogenouspathway.These results indicate that surgically

induced pancreatic growth is a sustainedphenomenon and that PBD can induce thedevelopment of premalignant microscopicfoci.

Ethanol and coeliac ganglionectomy in thedog

C D JOHNSON, M A DEVAUX, AND H SARLES(INSERM U3J, 46 Blvd de la Gaye,Marseille, France) Different doses ofethanol given during secretin infusion pro-duce either stimulation (low dose) or inhibi-tion (high dose) of pancreatic exocrinesecretion.' Low dose stimulation of proteinsecretion is abolished by pentolinium,but fluid secretion was not affected.'Pentolinium and truncal vagotomyabolished the high dose inhibition of bothfluid and protein output.2 To investigatethe neural pathways involved in theseresponses, the effect of celiac ganglionec-tomy (CG) was studied.During infusion of secretin 0.5 CU/kg/h,

ethanol diluted 10-20% in saline was givenas an intravenous infusion over 20 minutesto dogs prepared with gastric and duodenalThomas cannulas. Two doses were given inseparate experiments: 056 g/kg (controlsn=4; CG n=5) and 1 g/kg (controls n=6;CG n=4). Pancreatic fluid was collected in20 minute aliquots for two hours from thestart of the ethanol infusion. In controls,ethanol 0.56 g/kg stimulated pancreaticfluid (+0.041±0.008 ml/kg/min, sum of six

observations, mean±sem), and protein out-put (+0.26±0.08 mg/kg/min). After CG,ethanol 0.56 g/kg inhibited pancreatic secre-tion (fluid - 0-10±0-056 ml/kg/min, U=2,p=0-032; protein - 0-52±0-1 mg/kg/min,U=1, p=0.016). Ethanol 1 g/kg inhibitedpancreatic secretion in controls; CG had noeffect on this response.We conclude that the stimulation of pan-

creatic exocrine secretion by low dose intra-venous ethanol is mediated through theceliac ganglia and plexus. The inhibitoryresponse to high dose ethanol is mediatedby the vagus nerves and is unaffected byceliac ganglionectomy.

References1 Noel-Jorand MC, Sarles H. Dig Dis Sci 1983; 28:879-88.

2 Tiscornia OM, Hage G, Palasciano G, et al. Biomedicine1973; 18: 159-63.

Pancreatic pseudocysts in children

A P DICKSON, A J W MILLAR, H RODE, R JSTUNDEN, AND S CYWES Department ofPaediatric Surgery, Institute of ChildHealth, University of Cape Town and RedCross War Memorial Children's Hospital,South Africa) The management of 31 child-ren with pseudocysts of the pancreas isreviewed. The most common aetiology hasbeen trauma (58%) followed by ascarisworm infestation (20%). The frequentassociation with ascaris reflects its preval-ence in the local community (70%).

Clinical presentations include abdominalpain, vomiting, anorexia, fever andabdominal mass. Diagnoses have been con-firmed by persistent hyperamylasaemia,barium meal, and more recently ultrasoundand CT scans.Ten children (32%) were managed non-

operatively. Serum amylase returned tonormal at least one week before resolutionof the cyst. Spontaneous resolution hasoccured even when ERCP has shown a largecavity with duct disruption.

Fourteen patients (45%) were managedby cystgastrostomy for clinical indications.Operation was done either within a monthof presentation for an enlarging abdominalmass (seven patients), or delayed for anestablished or persistent lesser sac pseudo-cyst (seven patients).The only death in the series followed

cystgastrostomy anastomosis disruptionand septicaemia.Four patients (13%) underwent cystjeju-

nostomy without complication. In two suchpatients ERCP showed a major ductanomaly pre-operatively.



j.com/



ownloaded from

http://gut.bmj.com/

Pancreatic Society of Great Britain and Ireland

External drainage was used in threepatients, in two who had undergone pre-vious diagnostic laparotomy and in onewhose pseudocyst had ruptured.

Persistence of symptoms, signs andhyperamylasaemia for more than fourweeks indicated failed resolution in all butone case. Operation and internal drainagein unresolved cases was well tolerated. Inchildren whose fluid collections weremultiple or unstable ERCP identified pan-creatic duct anatomy.

Inhibition of postprandial pancreaticenzyme secretion by oral bile acid adminis-tration

A LANZINI, N HALL, D FACCHINETH, AND T CNORTHFIELD (Department of Medicine,St George's Hospital Medical School,London) Pancreatic proteases reduceserum CCK concentrations, inhibit post-prandial pancreatic enzyme secretion andpartially relieve chronic pancreatic pain.Ursodeoxycholic acid (UDCA) inhibits gallbladder emptying, possibly by reducingCCK concentrations, and might thereforeinhibit pancreatic secretion. We have there-fore studied the response of six gall stonesubjects to a liquid meal before and duringgall stone dissolution therapy with oralUDCA 675 mg daily. We used a duodenalperfusion technique incorporating PEG asa duodenal recovery marker, ICG as ahepatic bile marker and 99mTc HIDA as agall bladder bile marker (Lanzini, et al,Gastroenterology 1986; 92: 852-61). Duringthe first postprandial hour, the cumulativevalue for mean duodenal trypsin outputmeasured at 10 minute intervals wasreduced from 253 to 164 IU/kg/h (p<005),and the value for gall bladder ejectionfraction from 38 to 20% (p<0-05). Themean time taken to reach the peak valuewas delayed from 30 to 50 minutes fortrypsin output (NS) and from 25 to 45minutes for gall bladder ejection (p<0.05).There was no quantitative change in hepaticbile acid secretion rate during the first fivehours postprandially. We conclude thatUDCA attenuated the postprandialresponse to food for pancreatic trypsinsecretion as well as for gall bladder empty-ing (propbably because of the qualitativechange in biliary bile acid compositionknown to accompany UDCA therapy).UDCA may therefore have a role in thetreatment of chronic pancreatic pain, eitherby itself or in combination with pancreaticenzymes.

Chronic pancreatitis: a review of thesurgical pathology in 32 consecutive cases

M LAVELLE-JONES, M P HOLLEY, AND A

CUSCHIERI (Departments of Surgery andPathology, Ninewells Hospital and MedicalSchool, Dundee) Thirty two patients (18men, 14 women) aged 28-74 years under-went major pancreatic resection for chronicpancreatitis (CP). Of these, 14 underwenttotal pancreatectomy (four with segmentalpancreatic autotransplantation) and theremainder various lesser resections. Themajor aetiological factors were alcohol (15,10 men), trauma (4/32) and idiopathic (9, 8women).Although portal hypertension and

cirrhosis existed in only two individuals, 10had sectorial (left sided) portal hyperten-sion with splenomegaly (9/10) and splenicvein thrombosis (6/10). Severe extrapan-creatic fibrosis (EPF) occurred in 25/32(61%) patients. In 15/25, the lesser sac wascompletely obliterated. Pancreatic calcifica-tion (8) and sectorial portal hypertensionwere inevitably associated with EPF. 'Cysts'were present in 10/32 (30%), five wereintrapancreatic, three extrapancreatic andtwo were multiple.

Pancreatic ductal abnormalities onlyoccurred in individuals with EPF. Sevenducts were strictured, three had a chain oflakes appearance while the remainder werenormal. There was no generalised ductdilatation. Two splenic artery aneurysmswere noted, and the artery was extrinsicallycompressed and distorted (pseudo-aneurysmal) in two others. The commonbile duct was dilated in 7/32, and two ofthese had a pancreatic cancer. Overall, 3/32had pancreatic ductal adenocarcinoma and1/32 a pancreatic cystadenocarcinoma.

In conclusion (1) EPF and the absence ofmajor duct dilatation frequently renderspancreatic resection the only surgical optionin CP; (2) EPF rather than cirrhosis causessegmental portal hypertension in CP; (3)Intrapancreatic cysts are the commonestcystic abnormality in CP; (4) EPF limits theapplication of segmental pancreatic auto-transplantation techniques in patients withCP.

Role of duodenal bile crystal analysis in theinvestigation of patients with 'Idiopathic'pancreatitis

J P NEOPTOLEMOS, B R DAVIDSON, D VALLANCE,AND A F WINDER(Department of Surgery and ChemicalPathology, Leicester Royal Infirmary,

Leicester) The aim of this study was to assessthe diagnostic value of gall bladder bilecrystal analysis (cholesterol monohydrateand calcium bilirubinate crystals) in patientsclassified as 'idiopathic' pancreatitis afterinvestigation by ultrasonography andERCP. As a corollary the types of gallstones in patients with acute pancreatitiswere also analysed.

Bile was collected at duodenoscopy after100 U iv CCK-PZ (Boots). Samples wereanalysed independently by direct andpolarising microscopy at 20°C and afterincubation at 37°C for 24 hours. Gall stoneswere analysed by infrared spectroscopyusing a Perkin-Elmer 598 spectrophoto-meter.The technique was first assessed in 34

patients without pancreatitis (21 with gallstones and 16 without) and in 26 with acutepancreatitis of known aetiology (11 causedby gall stones and 15 caused by alcohol).The sensitivity was 64%-67% and thespecificity was 94%-100% in these groups.Calcium bilirubinate crystals (but notcholesterol crystals) were found in five of 14(36%) patients with 'idiopathic' pan-creatitis. Gall stones were found in four ofthese patients at cholecystectomy (29%),three of whom were positive for crystals.Pigment stones were found in nine of 13(69%) pancreatitis patients compared toseven of 31 (23%) controls (p<0.005).The finding of a high incidence ofpigment

stones in patients with pancreatitis indicatesthe importance of including calcium biliru-binate crystal analysis (as well as cholesterolcrystals) in the assessment of duodenal bile.These results suggest that duodenal bilecrystal analysis may be a useful techniquefor the investigation of patients with acutepancreatitis.

Multivariate analaysis of prognosticmarkers in acute pancreatitis: can pan-creatic collections be predicted?

T LEESE, W M THOMAS, D SHAW, M FULLER,AND M HOLLIDAY (Departments of Surgeryand Pathology, University of Leicester) Inpatients with acute pancreatitis it has beensuggested that the combined determinationof serum C-reactive protein and alpha2macroglobulin may replace contrastenhanced computed tomography in theearly detection of pancreatic necrosis. 'The value of six prognostic markers was

assessed prospectively in 200 attacks ofacute pancreatitis with specific attention totheir ability to predict pancreatic collec-tions. The Imrie Prognostic Score (IPS) was

A268



j.com/



ownloaded from

http://gut.bmj.com/


recorded within 48 hours of diagnosis.Serum C-reactive protein (CRP), alpha,antiprotease (AlAP), alpha2 macroglobu-lin (A2M), amylase and white cell count(WCC) were measured on days 1, 3, and 7.When comparing all severe clinical out-

comes to mild outcomes serum CRP con-centrations were higher on all three days(p<0.02, <0-001, <0.001), AIAP concen-trations were higher on day 3 (p<005),A2M concentrations were lower on day 7(p<0.01) and WCC was higher on all threedays (p<0-001, <0-001, <0.001). Serumamylase concentrations showed no signifi-cant differences. None of the measuredparameters were helpful in distinguishingpatients who subsequently developed pan-creatic collections from patients who hadother severe outcomes.

Multivariate analysis revealed that theinitial IPS showed greatest independentsignificance in predicting severe outcomefollowed by the WCC (days 1 and 7) andCRP (day 3).Day three serum CRP >0-15 g/l and day

seven WCC >15x109/l may be clinicallyuseful predictors of severe outcome tosupplement the initial IPS. These methodsare unlikely to distinguish pancreatic collec-tions from other severe outcomes, but maysupplement clinical judgement in selecting ahigh risk group of patients for contrastenhanced computed tomography.

Reference1 Buchler M, Malfertheiner P, Schoetensack C, Uhl W,Scherbaum W, Beger HG. Wertigkeit biochemischerund bildgebender verfahren fur diagnose und prognosedes akuten pankreatitis. Z Gastroenterol 1986; 24:100-9.

Value of C-reactive protein and anti-proteases in the objective monitoring ofacute pancreatitis

C WILSON, A SHENKIN, AND C W IMRIE(Division of Surgery and University Depart-ment of Biochemistry, Royal Infirmary,Glasgow) Prediction of severity in acutepancreatitis by multiple factor scoring ordiagnostic peritoneal lavage identifies agroup of patients at increased risk of deathor early complication. Neither permitsobjective day-to-day monitoring during theillness which might usefully assess theresponse to therapy or warn of a latecomplication. We have investigated alpha-2-macroglobulin (a2M), alpha-i-anti-protease (alAP) and C-reactive protein(CRP) in 72 patients to determine the valueof their sequential determination.A2M concentrations fell during the ill-

ness with concentrations significantly lowerin complicated cases from day 3 to 8.C-reactive protein and alAP showed amarked rise during the course of the illness,aIAP concentrations being significantlyhigher in complicated cases during days 4 to8. C-reactive protein showed the best discrim-ination, the differences between mild andcomplicated attacks being highly significant(p<0001) from day 2 to 8. C-reactive proteindid not distinguish patients dying withinthree days of admission. Eight of 1 1 patientsdeveloping necrosis or pseudocyst had apeak CRP >300 mg/1 but the mean CRPcould not distinguish these patients, six ofwhom required surgery, from nine patientswith acute respiratory failure who settledwithout surgery. Nevertheless sequentialmeasurement of CRP promises to be aclinically useful advance, permittingobjective monitoring of the course of anattack.

Plasma exchange for acute necrotising pan-creatitis. A preliminary report

M C PARKER, S CAWTHORN, M J ALLEN, E DBENNErr, AND J HERMON-TAYLOR (Depart-ments of Surgery and Intensive Care, StGeorge's Hospital, London) The mortalityof non-operative treatment of acute necro-tising pancreatitis (ANP) is reported as upto 100% whereas that for surgery variesbetween 22% and 64%. This suggests thatthe ideal treatment for ANP is surgical. Theselection of patients suitable for surgery ismade more difficult, however, by the lack ofa prognostic indicator. Such patients usuallydevelop multi-organ failure (MOF), bywhich time they are often considered unfitfor surgery. MOF is thought to be caused bycirculating activated proteolytic enzymesfrom the damaged pancreas with sub-sequent activation of inflammatorymediators. Although many of thesemediators are membrane bound, thosewhich circulate freely or are protein boundare theoretically removable by plasmaexchange.We have performed plasma exchange on

five occasions in four patients, each with aminimum of three organ system failure, asthe result of ANP. There were no adverseeffects attributable to the treatment. Two ofthe four patients showed improvement inmean arterial blood pressure and urineoutput. One of these subsequently under-went pancreatic resection and survived.Surgery was refused in the other, whounderwent a second plasma exchange,again with improvement but who eventually

died. Both patients who failed to respond toplasma exchange died, one within two hoursof treatment and the other 24 hours postlaparotomy. At operation pancreatic resec-tion had been deemed impossible.We conclude that plasma exchange may

be useful in the immediate preoperativeperiod in patients with ANP selected foremergency pancreatic resection.

Impaired clearance of circulating proteaseantiprotease complexes in severe acutepancreatitis: an important aspect ofpathogenesis?

M LARVIN, S F SWITALA, AND M J MCMAHON(University Department of Surgery, TheGeneral Infirmary at Leeds, Leeds) Trypsinliberated into the circulation is rapidly andpreferentially complexed by alpha2 macro-globulin (oC2-M).' Trypsin- oC2-M com-plexes retain peptidase activity, andimbalance of complexed:free oc2-M leads tofatal shock in experimental animals.2 Inhealthy individuals circulating trypsin-oc2-M complexes are quickly cleared by thereticulo-endothelial system (RES), with ahalf-life of 8-10 minutes.' In patients withsevere acute pancreatitis (AP), up to 30%of plasma oc2-M is complexed, suggestinggrossly impaired clearance.3 The clearanceof circulating trypsin- oC2-M complexes wasstudied in four healthy persons, and 17patients with AP (11 mild, six severe,according to clinical criteria) within 24hours of admission to hospital. After with-drawing a baseline blood sample, 1 mg 'C-trypsin (specific activity 370 kBq/mg) wasinjected intravenously. Blood samples werecollected immediately, at 1 minute intervalsfor 20 minutes, and at 2 minute intervals fora further 20 minutes.Sample radioactivity was counted by

liquid scintillation, and the molecular dis-tribution of the radiolabel verified bySephadex G200 gel filtration. Plasma '4C-trypsin- °'2-M half life (tl,2) was calculatedas the time to 50% reduction of initialactivity.

Plasma 'C-trypsin- xC2-M t1l2 (minutes)n median range (*Two-sample rank test)

Normal 4 11-1 8-1-13-2Mild 11 12-5 4-8-26.8 p<0-02*Severe 6 21-6 17-7-31.0 J p<O.05*

The study shows impaired clearance ofcirculating protease-antiprotease com-plexes during AP. In patients suffering fromsevere AP, the continuing biologicalactivity of the complexes may justify



j.com/



ownloaded from

http://gut.bmj.com/


plasmapheresis or plasma exchangetherapy.

References1 Ohisson K, Laurell C-B. The disappearance of enzyme

inhibitor complexes from the circulation of man. Clin SciMol Med 1976; 51: 87-92.

2 Ohlsson K. Experimental pancreatitis in the dog.Appearance of complexes between proteases and tryp-sin inhibitors in ascitic fluid, lymph and plasma. ScandJGastroenterol 1971; 6: 645-52.

3 Lasson A, Ohisson K. Protease inhibitors in acutehuman pancreatitis. Correlation between biochemicalchanges and clinical course. ScandJ Gastroenterol 1984;19: 779-86.

Exchange transfusion reduces the mortalityin experimental pancreatitis in the rat

A L BLOWER AND T V TAYLOR (Department ofSurgical Gastroenterology, ManchesterRoyal Infirmary, Manchester) No specifictreatment exists for acute haemorrhagicpancreatitis (AHP), for which the mortalityis 75%. Exchange transfusion (XT), whichoffers a means of clearing pancreaticenzymes and immune complexes from thecirculation and may supplement depletedstocks of enzyme inhibitors, has beeninvestigated.

In the male Sprague-Dawley rat, AHPwas induced by infusion of a bile-salt/lysolecithin mixture into the bile-pancreaticduct. There was rapid development ofblood stained ascites and rise in serumamylase.

Fifteen millilitres donor blood from twodonor animals was added to 3 ml acid-citrate-dextrose. Exchange transfusion(XT), performed via tail vein and femoralartery cannulae, involved the removal andreplacement of 1 ml blood on 18 separatebut continuous occasions. An exchangeefficiency of 70% was obtained as assessedwith radio-labelled serum albumin. Allanimals were monitored until death orkilled at 72 hours.

In 20 untreated animals the mortality at24 hours was 60% and at 72 hours was 90%.In 20 animals XT at 1 hour reduced themortality to 10% and 20% respectively(x2= 1099, p<-001 and x2= 19-8,p<O0OOOl. The 72 hour mortality if XT wascarried out at 0-5, 1, 1-5, 2, and 4 hours42%, 50%, 67%, and 87% respectively(n/s). At 30 minutes, however, the pan-creatitis was not fully developed histologic-ally and after one hour it may have been toolate in the disease process to affect theoutcome. The administration of freshplasma 1 hour after induction of AHP didnot reduce the mortality.Exchange transfusion reduced the 72

hour mortality from experimental AHP by

70%. The timing of XT appears to be ofgreat importance.

Intraperitoneal antiprotease therapy inexperimental acute pancreatitis

C WILSON, G D MURRAY, C W IMRIE, AND D CCARTER (Department of Surgery, RoyalInfirmary, Glasgow) Experimental worksuggests that the peritoneal antiproteasedefences may be overwhelmed in canineacute pancreatitis, leading to shock anddeath of the animal. Furthermore, highdegrees of saturation of peritoneal alpha-2-macroglobulin have been demonstrated insevere acute pancreatitis in man. Removalof the peritoneal pancreatitis exudate andreplacement with fresh antiproteases mightthen be therapeutic. This concept has beeninvestigated in an experimental pancreatitismodel.

Fifty male Wistar rats had pancreatitisinduced by intraductal infusion of 5%

sodium taurocholate and, before peritonealclosure, a cannula was placed in the peri-toneal cavity. One hour later 40 animals(comprising four groups of 10) underwentaspiration of the peritoneal exudate and asingle peritoneal lavage with 0-9% saline(1.5 ml/100 gm body weight). These ratsthen had instillation into the peritonealcavity of an anti protease or a similarinactive solution (1.5 ml/100 g bodyweight):0-9% saline, aprotinin (5000KIU/ml saline), human plasma proteinsolution or human fresh frozen plasma. Theremaining 10 rats had no intraperitonealtherapy (controls) having instead 0.9%saline (1.5 ml/100 g body weight) injectedsubcutaneously at one hour. Peritonealcannulas were then removed and the sur-vival times monitored (Table).

In all animals having peritoneal therapymedian survival time increased compared tothe control group (I2=1006, 4 df, p=0.04).Instillation of fresh frozen plasma appearedthe most effective of the intraperitonealtherapies tested.

Median Interquartile p value comparedGroup n survival (h) range (h) with controls

Control 10 9-5 8-8-13-4 -

Saline 10 20-3 7-9-51-4 0-236Plasma protein sol. 10 17-5 14.9-22.5 0-116Fresh frozen plasma 10 24-8 18-8-72 0-01*Aprotinin 10 19-8 14-0-23-3 0-065

*Significant, log rank test.

Depressed mononuclear phagocyte functionin severe acute pancreatitis

S F SWITALA, M LARVIN, AND M J MCHAHON(University Department of Surgery, TheGeneral Infirmary, Leeds) Suppression ofphagocytic cells during acute pancreatitis(AP) in experimental animals increasesmortality,' whilst immunostimulating drugsdecrease mortality.'2 Impaired function ofphagocytic cells has been demonstratedin malignant disease and septic states.Proteases released into the circulationduring AP are complexed by X2-macroglobulin and cleared by mononuclearphagocytes. The aim of the present studywas to investigate the relationship of

phagocytic cell function to the outcome ofacute pancreatitis. Polymorphonuclearneutrophil and mononuclear phagocyteswere isolated from peripheral blood bya modification of Boyum's method, andcultured separately in monolayers. Afterstimulation with zymosan-A (5 x 106particles per 1x 106 cells), monolayers wereexamined microscopically. The phagocyticindex (%PI) was calculated as the propor-tion of cells containing two or moreparticles. The subjects of the study werehealthy controls (n=8) and patients withAP, studied within 24 hours of admissionto hospital (n= 18, 12 mild and 6 severeattacks).Monocyte phagocytosis was significantly

Monocyte %PI Polymorph %PI

n median range median range

Controls 8 27-0 18-3-77-0 T 358 14-5-91-0 TMild 12 20-2 11-8-59-0 p<0-05* 40-7 9-2-82-6 NS*Severe 6 9.2 2-2-36-4 -4 35-2 17.2-47.1 4.

*Two-sample rank test.



j.com/



ownloaded from

http://gut.bmj.com/


decreased in patients with severe AP com-pared to the control group, although PMNphagocytosis was not significantly changed.Depressed mononuclear phagocyte func-tion may be deleterious in AP, as it mayresult in impaired clearance of biologicallyactive protease-antiprotease complexes.

References1 Adham NF, Song MK, Haberfield GC. Relationshipbetween the functional status of the reticulo-endothelialsystem and the outcome of experimentally inducedpancreatitis in young mice. Gastroenterology 1983; 84:461-9.

2 Browder LW, Sherwood E, Williams D, Jones E,McNamee R, DiLuzio N. Protective effect of Glucan-enhanced macrophage function in experimental pan-creatitis. Am J Surg 1987; 153: 25-33.

Successful xenogeneic islet transplantationin the diabetic nude (athymic) rat

S P LAKE, J CHAMBERLAIN, P HUSKEN, P R FBELL, AND R F L JAMES (Department ofSurgery, University of Leicester, Leicester)Transplantation of isolated human pan-creatic islets offers a potential treatment fordiabetes, however the isolation process ismultistaged and potentially damaging.Current methods of assessing islet viabilityutilise in vitro techniques - for example,vital stains and measurement of insulinrelease. A method of assessing in vivoviability would be a useful adjunct in isletisolation and storage techniques.The nude (athymic) rat offers such a

potential model, however the induction ofdiabetes with streptozotocin results in ahigh mortality, mainly because of dehydra-tion during the two week postinductionperiod. In this study we have transplantednude rats with xenogeneic (mouse) islets 48hours post-induction of diabetes, and tocontrol for possible reversion by therecipient we have used the renal sub-capsular space to allow for subsequent graftremoval.

Fifteen nude rats were given streptozo-tocin (70 mg/kg) intravenously and 14became diabetic (blood glucose >20 mmol/1) within 48 hours. Seven control animalslost weight (0.5±0.2 g/day) and remaineddiabetic until death (mean survival, 23±1-8days). Diabetes was reversed within 48hours in 7/7 animals transplanted with 500-750 mouse islets. These animals remainednormoglycaemic, gained weight (mean1-9±0-3 g/day) and during glucose toler-ance testing, at 21 days, showed nearnormal glucose response. Graft removalresulted in reversion back to the diabeticstate in all animals. Histology of theresected grafts confirmed the presence ofislet tissue with no evidence of rejection.

Diabetes has been successfully reversedin the diabetic nude rat by xenogeneic islettransplantation. This model offers a poten-tial method for assessing in vivo function ofisolated human islets.

Cyclosporin A allows prolonged canine isletallotransplant survival

A J GUY, A M EL-SWAIS, S M GRIFFIN, DALDERSON, AND J R FARNDON (Department ofSurgery, The University of Newcastle,Newcastle-upon-Tyne) To date islet allo-transplants have been largely unsuccessfulin the reversal of experimental diabetes inlarge animals. The effect of Cyclosporin A(CsA) on islet auto and allo-transplantsurvival was investigated using 27 pan-createctomised beagles in four groups;autotransplants, immunosuppressed auto-transplants, allotransplants, iminuno-suppressed allotransplants. Single donorpancreatic microfragment grafts were pre-pared by collagenase perfusion' and trans-planted into the splenic pulp. CsA was givenintramuscularly to dogs in groups 2 and 4 atan initial dose of 20 mg/kg/day, adjusted tomaintain the plasma CsA concentrationabove 400 ng/ml as measured by radio-immunoassay. Plasma glucose and insulinwere measured in the post-transplantperiod.Plasma glucose was significantly higher

for group 2 animals than group 1 on post-transplant days 4, 6, 7, and 28 (p<0-02,Mann-Whitney-U test), but plasma insulinconcentrations were no different. At threemonths one of six grafts had failed in group1, and three of six in group 2 (NS, Fisher'sexact test). CsA treatment significantly pro-longed islet allotransplant normoglycaemicsurvival (group 3 v group 4, p<0-005,Mann-Whitney U-test). Six long term sur-vivors from groups 1, 2, and 4 underwentsplenectomy, removing the islet graft: theyall rapidly became severely diabetic,demonstrating unequivocal graft depend-ence.

In conclusion, the use of Cyclosporin A asthe sole immunosuppressant, monitored tomaintain an adequate plasma CsA concen-tration, has allowed significant prolonga-tion of canine islet allotransplant function.

Reference1 Griffin SM, Alderson D, Farndon JR. Comparison of

harvesting methods for islet transplantation. Br J Surg1986; 73: 712-5.

Pancreatic islet transplantation: endocrinefunction and effect on secondary diabeticcomplications

T N WALSH, D ALDERSON, AND J R FARNDON(Departments of surgery, MaterMisericordiae Hospital, Dublin, RoyalVictoria Infirmary, Newcastle-upon-Tyne)Conventional insulin therapy does not pre-vent the secondary complications of insulin-dependent diabetes probably because per-sistent metabolic derangement has beenimplicated in their pathogenesis. This studyexamined diurnal insulin and glucose pro-files in a large animal model of insulin-dependent diabetes (pancreatectomiseddog) treated with insulin or pancreaticmicrofragment autotransplantation. Glyco-sylated haemoglobin levels and fluoresceinangiography were carried out at 18-27months folldwing pancreatic autotrans-plantation. Three groups of dogs werestudied. Group 1 - 10 normal controls;group 2 - six pancreatectomised dogs withintrasplenic dispersed pancreatic isletmicrofragment autografts and group 3 -seven pancreatectomised insulin treateddiabetic dogs. At two weeks (group 3) andthree months (group 2) insulin and glucosediurnal profiles were examined over a16-hour period from 0600 h to 2200 h.Glycosylated haemoglobin levels weredetermined monthly. At 18 and 27 monthsfluorescein angiography was performed inanimals with successful autotransplant andcompared with normal dogs. Islet trans-plant recipients had insulin profiles identicalto normal but diabetic animals showedgrossly raised insulin concentrations post-prandially and the concentrations wereabnormally low in the fasting state. Despitemild carbohydrate intolerance, trans-planted dogs had normal glycosylatedhaemoglobin concentrations (mean 6-88+0-86, normal range 6-86±0-5). At 18 and 27months transplanted animals showed noevidence of diabetic retinopathy on fluores-cein angiography. By restoring normal

Results

Group 1 (n=9) Group2(n=6) Group3(n=6) Group4(n=6)autotx. autotx. +CsA allotx. allotx. +CsA

Days of >28x3,91, 26,42,49, 2,3,4, 22,35,41,normoglycaemia >182x5 >84x3 4,5,5 56,64, >182x 1



j.com/



ownloaded from

http://gut.bmj.com/


insulin and glycaemic profiles pancreaticislet autotransplantation prevented thesecondary complications of diabetes andmight represent the ideal therapy for humaninsulin-dependent diabetes.

Longterm metabolic follow up using differ-ent preparations of islet autografts

S M GRIFFIN, D ALDERSON, AND J R FARNDON

(Department of Surgery, The University ofNewcastle-upon-Tyne, Newcastle) Largeanimal islet autografts restore diabeticmetabolism to normal in the short term.Complications of diabetes appear only aftera significant duration of disease process. Asassessment of longterm function of graftsneed to be made before allograft experi-ments are entertained. Twelve dogs were

made diabetic by total pancreatectomy.Seven dogs received 50% of a graft pre-pared by an established technique (group1). Five dogs received 33% of a graftprepared by a method previously shown tobe initially effective (group 2).' Tissue wasgrafted into the splenic pulp in all animals.At 1, 3, 6, 12, 18, 24, and 30 monthseuglycaemic animals were assessed byIVGTT so that glucose clearance, insulin,lactate, pyruvate, alanine, FFA, glycerol, 3-hydroxybutyrate and glucagon responsescould be determined. Splenectomy was per-formed on one animal in each group at twomonths to confirm graft dependence.

All animals in group 1 became diabetic byeight months, with fasting hyperglycaemiaoccurring at 2, 4, 4, 7, 8, and 8 monthsrespectively. No animals in group 2 becamespontaneously diabetic. Abnormalities infat metabolism were noted up to three

months had normalised by six months andmetabolic profiles were normal up to 30months. Histology of the spleens confirmedthe presence of the three islet cell types.Delayed graft failure does occur using theestablished technique of graft preparation(group 1), but long term function can beachieved using the group 2 method. Inaddition, metabolic control remains normalfor up to 30 months.

Reference1 Griffin SM, Alderson D, Farndon JR. Comparison of

harvesting methods for islet transplantation. Br J Surg1986; 73: 712-5.

Enquiries regarding the Society should beaddressed to the Secretary:Mr CW Imrie,Division of Surgery,Royal Infirmary,Glasgow G4 OSF.



j.com/



ownloaded from

http://gut.bmj.com/

Documents

PancreaticSociety ofGreat Britain andIreland · patients undergoing surgery for chronic pancreatitis. Thehistology has beenevalu-ated by an experienced liver histopatholo-gist and