Abstracts from the Twenty-Fourth Meeting of the Pancreatic Society of Great Britain and Ireland at the Royal College of Physicians, London, UK November 23–24, 1999

International Journal of Pancreatology

259

Volume 27, 2000

International Journal of Pancreatology, vol. 27, no. 3, 261–267, June 2000 © Copyright 2000 by Humana Press Inc. All rights of any nature whatsoever reserved. 0169-4197/00/27:261–267/$11.75

Oral Presentations

There were 172 patients with acute pancreatitis recruited, of whom 35 had severe disease as defined by the Atlanta criteria. Seventy-four patients with acute abdominal pain of other causes served as controls. TAP levels were measured in urine samples taken on admission and at 12, 24, 36, 48, 72, 96, and 168 h thereafter (TAPKIT, Biotrin International Ltd., Dublin, Ireland). Urinary TAP peaked 12 h after symptom onset at 111 ± 45 nM (mean ± SEM) in patients with severe disease compared with 36 ± 11 nM for mild disease and 20 ± 7 nM for controls ( p < 0.05). Urinary TAP levels of >35 nM within 24 h of symptom onset differentiated severe from mild disease with sensitivity (S), specificity (Sp), and negative predictive values (NPV) of 58, 73, and 86%, respectively. This compares with 0, 90, and 75% for CRP (>150 mg/L). At 24 h postsymptom onset the area (95% CI) under the receiver operator characteristic curve was greater for TAP (0.69; 0.51–0.88) than for CRP (0.40; 0.23–0.58). Urinary TAP levels remained consistently higher in patients with severe acute pancreatitis up to 84 h postsymptom onset. At 48 h posthospitalization the S, Sp, and NPV for TAP (>35 nM) were 83, 72, and 94%, compared with 86, 61, and 94% for CRP (>150 mg/L), 56, 64, and 85% for APACHE II (>8), 89, 64, and 96% for Ranson (≥3) and 77, 75, and 93% for Imrie (≥3). Urinary TAP is at least as good a predictor of severity as any of the currently utilized scoring systems, but its simplicity and applicability within 12 h of symptom onset makes it an extremely valuable severity marker in the assessment of patients with acute pancreatitis.

A Human Model of Acute Pancreatitis: Premature Protease Activation In Human Pancreatic Explants J B Ockrim, M L Haw, F Compton, A C Steger, G M Murphy, M L Wilkinson Gastroenterology, GKT (St Thomas’ Campus), King’s College London Background. A human model for acute pancreatitis was derived using pancreatic biopsies that were obtained after informed consent from patients undergoing a pancreaticoduodenectomy. Fragments of the tissue were suspended in culture medium prior to storage in liquid nitrogen. A cerulein dose response study was performed on thawed aliquots of these human pancreatic explants. Results. Histological examination showed normal pancreatic architecture. Trypan blue exclusion was more than 95%. Amylase secretion increased from 520 IU/µg cell protein in the absence of cerulein to a peak of 1270 IU/µg cell protein at 0.1 nM cerulein and thereafter fell with increasing cerulein concentrations. Trypsinogen secretion mirrored that of amylase rising from 0.275 µg cell protein in the unstimulated cell to a peak of 2.94 IU/µg cell protein at 0.1 nM cerulein. Trypsin secretion was detected only at the highest cerulein concentration (74 nM) when the amylase level was 1001 IU and the trypsinogen 2.65 IU/µg cell protein. Conclusion. We have shown for the first time that viable human pancreatic acinar explants may be readily obtained from small amounts of resected tissue and that this model can be used to study acute pancreatitis.

Altered Intestinal Morphology and Depletion of Intestinal Immune Cells in Patients with Necrotizing Acute Pancreatitis

Urinary Trypsinogen Activation Peptide is a Reliable Early Predictor of Severity in Acute Pancreatitis

B J Ammori, A Cairns*, M F Dixon*, M Larvin, M J McMahon

M G T Raraty1, J P Neoptolemos1, J Slavin1, E Kemppainen2, A Hieteranta2, P Puolakkainen2, J Fitzpatrick3, J Mayer 4, H-G Beger4

The Academic Surgical Unit and *Department of Pathology, The General Infirmary at Leeds, UK

Depts. of Surgery, University of Liverpool, UK1, University of Helsinki, Finland2, Mater Hospital, Dublin, Ireland3 and Ulm University Hospital, Germany4.

Background. Impairment gut barrier function has been demonstrated in patients with severe acute pancreatitis and may contribute to the development of local and systemic septic complications. The underlying mechanisms, however, remain unclear. Aims. To investigate the small intestinal morphology and intestinal mast cell content (immune cells) in patients severe acute pancreatitis.

There is an urgent need for a reliable early prognostic marker in acute pancreatitis. We report the results of a prospective multicenter study comparing a new competitive immunoassay for tripsinogen activation peptide (TAP) with three standard clinico-biochemical scoring systems and with C-reactive protein (CRP).

261

262

Abstracts

Methods. Segments of terminal ileum from three patients with severe necrotizing acute pancreatitis who underwent pancreatic debridement and ileocolic resection for doubtful or evident segmental colonic viability were available for the study. Control specimens were available from seven patients who underwent gastric bypass and distal ileal resection for morbid obesity. Sections were cut and stained with haematoxylin and eosin for the measurement of villous height and crypt depth, and with toluidine blue for the determination of mucosal mast cell counts. Only adequately oriented specimens were deemed suitable for computer-aided image analysis using a Leitz image analyzer system. Results were expressed as villous height/crypt depth ratios (VH/CD) and mucosal mast cell index (ratio of mast cell count/length of muscularis propria). Results. Microscopy of the small intestine from controls was normal. The villous height and the VH/CD ratio were significantly reduced in patients with acute pancreatitis compared with controls (median 0.47 mm vs. 0.68 mm, and 1.9 vs. 2.8, respectively, p < 0.0001). The mast cell index was significantly reduced in patients with acute pancreatitis compared with controls (median 5.88 cell/mm vs. 8.58 cell/mm, p = 0.001). A positive correlation was observed between the mast cell index and the height of the villi (r = 0.23, p = 0.027). Conclusion. Patients with necrotizing acute pancreatitis have an altered intestinal morphology and depleted mucosal mast cells. These factors that may contribute to the impairment of gut barrier function in patients with severe acute pancreatitis.

Evidence of a Genetic Predisposition for Acute Pancreatitis A Buter, J Eskdale, G Gallagher, C R Carter, C W Imrie, C J McKay Dept of Surgery, Glasgow Royal Infirmary, Glasgow. Background. Why do some people develop acute pancreatitis and others not, when exposed to the same etiological factors? Why do some have a severe course once they have acute pancreatitis? These are important fundamental questions that remain unanswered. The tumor necrosis factor (TNF) gene cluster in the central major histocompatibility complex on chromosome 6 contains several polymorphisms, including the TNFα microsatellite. The TNFα2 allele is associated with a higher TNF production from lipopolysaccharide-stimulated monocytes in vitro. Because of the central role of TNF in the development of acute pancreatitis and its complications, we hypothesised that the TNFα2 allele is associated with acute pancreatitis and its complications. Methods. Genomic DNA was extracted from ethylenediaminetetraacetic acid blood samples of patients with clinical and biochemical acute pancreatitis. Following polymerase chain reaction amplification, the TNFα microsatellite alleles were resolved on denaturing polyacrylamide gels. The frequency of the TNFα2 allele was compared to the observed frequency in a

International Journal of Pancreatology

normal control group of healthy, nonrelated donors from the same West of Scotland population (n = 91). Results. One hundred and four patients, 56 with severe and 48 with mild, pancreatitis were included. The etiology was gallstones (67%), alcohol (16.5%), endoscopic retrograde cholangiopancreatography (3.3%), and unknown (13.2%). The TNFα2 allele was significantly more common in acute pancreatitis patients, than in normal controls (36.1 vs. 20.3%; p = 0.0006 chi square). There was no significant difference in the frequency of the TNFα2 allele between the mild and severe groups. Conclusions. This study has demonstrated for the first time a significant association of the TNFα2 allele with acute pancreatitis. A genetic predisposition may explain, at least in part, why some develop acute pancreatitis and others do not.

Mutations in the Cationic Trypsinogen and Cystic Fibrosis Transmembrane Regulator Genes in Patients with Idiopathic Chronic Pancreatitis J E Creighton, *C Roberts, *R Plunkett, *R Lyall, *D I Wilson, *A Curtis, R M Charnley Department of Hepato-Pancreatico-Biliary Surgery, Freeman Hospital, Newcastle, *Dept. of Human Genetics, Northern Region Genetics Service, Newcastle, UK Background. In up to 30% of cases of chronic pancreatitis, no underlying cause is found despite intensive investigation. Mutations in the cationic trypsinogen gene (PRSSI) have been shown to cause hereditary pancreatitis, the clinical features of which are identical to those of other forms of pancreatitis. Recent studies have also suggested that heterozygous mutations in the cystic fibrosis transmembrane regulator gene (CFTR) are associated with idiopathic chronic pancreatitis (ICP). This study investigates the frequency of mutations in these two genes in patients referred to a pancreatic clinic with a presumed diagnosis of ICP. Methods. Forty-eight consecutive patients attending a pancreatic clinic between 1/11/97 and 31/12/98 with a diagnosis of idiopathic chronic pancreatitis were studied. A further four patients referred with a presumed diagnosis of hereditary pancreatitis were excluded. Venous blood was taken to identify the 2 known mutations in PRSSI, and to screen for the 16 most common mutations in CFTR (accounting for 85% of cystic fibrosis alleles in this region). In addition, single-strand conformation polymorphism analysis of individual exons of CFTR, with sequencing of any abnormal result, was carried out to screen for rare mutations. Results. Nine (19%) patients referred with ICP were found to have mutations in PRSSI, consistent with a diagnosis of hereditary pancreatitis. A further 5 (10%) were shown to be heterozygous for CFTR mutations with one patient carrying two mutations (∆F508 and P67L). The population frequency of carriage of a CFTR mutation was 4.8% (n = 165), p = 0.05.

Volume 27, 2000

Oral Presentations Conclusions. Genetic abnormalities are common in patients referred to a specialist clinic with idiopathic chronic pancreatitis. Twenty-nine percent of patients seen over a 14-mo period have subsequently been shown to carry either a disease causing mutation in PRSSI, or to be heterozygous for CFTR mutations. The frequency of “mild” CFTR mutations (R117H and P67L) appears to be particularly increased. More mutations are likely to be found in CFTR if more extensive analysis is carried out. It is important to consider the possibility of mutations in PRSSI or CFTR in patients presenting with idiopathic chronic pancreatitis.

Hereditary Pancreatitis and the Risk of Pancreatic Cancer N Howes, M M Lerch, V Keim, J Mossner, S Endres, J Deviere, G Verreman, V Lucidi, R Charnley, C Imrie, A Olah, I Ihse, W Steenbergan, M O’Donnell, W Greenhalf, I Ellis, S Rutherford, R Mountford, D C Whitcomb, J P Neoptolemos for EUROPAC Department of Surgery, Clinical and Molecular Genetics, Liverpool on behalf of the consortium of EUROPAC Introduction. Hereditary pancreatitis (HP) carries a significant lifetime risk for the development of pancreatic cancer (PC). Aims. To quantify the risk of pancreatic cancer in HP families referred to the European Registry of Hereditary Pancreatitis and Familial Pancreatic Cancer (EUROPAC). Methods. Recruitment started in 1997. Families were considered if two members had chronic pancreatitis of unknown etiology. Families were tested for protease serine 1 (PRSS1) mutations using polymerase chain reaction restriction digestion, sequencing the PRSS1 gene screened negative families. Clinical information about chronic pancreatitis (CP) and development of PC was obtained from family members and referring clinicians. We compared the observed and expected frequency of pancreatic cancer in our historical cohort and estimated the association of HP and PC using the standardized incidence ratio (SIR), which is the ratio of observed PC to expected PC in the background population adjusted for age, sex, and nationality. Results. Information was obtained on 54 families (n = 136) comprising 72 females and 64 males. Nineteen families (n = 53) had the Rl17H mutation, 12 (n = 34) the N21I mutation and 20 (n = 37) neither mutation. In 3 (n = 12) families the result of mutation analysis is outstanding. Eleven patients (5 male, 6 female) mean age (± standard error) of 59 ± 3.0 yr developed PC during 5071 person-years, compared to an expected 0.12. This yielded an SIR (95% CI) of 93 (25–161). The estimated cumulative risk (95% CI) of PC to age 73 was 56% (17–73%). The mutation status of the patients with PC were 5 with Rl17H, 2 with N21I, and 3 were negative. No mutation information was available on one patient. Conclusion: HP carries a significantly increased risk of PC in our cohort of patients, which equates to an estimated life time risk for the development of PC of 56%.

International Journal of Pancreatology

263 Subcellular Localization of Enzyme Activation After Hyperstimulation in Mouse Pancreatic Acinar Cells M G T Raraty1,2, J P Neoptolemos1, R Sutton1, O H Petersen2 Departments of Surgery1 and Physiology2, University of Liverpool, Liverpool, UK We have previously demonstrated that intracellular activation of digestive enzymes occurs shortly after hyperstimulation of mouse pancreatic acinar cells with cholecystokinin (CCK) and after inhibition of calcium reuptake with thapsigargin. We report further experiments which demonstrate the sub-cellular localization of such enzyme activation. Isolated mouse pancreatic acinar cells were loaded with 10 µM 7-amino-4-chloromethylcoumarin, CBZ-L-isoleucyl-Lprolyl-L-arginine amide hydrochloride, a fluorescent substrate specific for trypsin. Fluorescence was visualized on a Zeiss confocal microscope using excitation and emission wavelengths of 364 and 466 nm respectively. All experiments were repeated at least six times. Approximately 300 s after hyperstimulation with l0 nM CCK, fluorescence was seen to appear within multiple discrete compartments within the granular region at the apical pole of the cell. A similar rise in fluorescence was observed after treatment with 2 µM thapsigargin. Pretreating the cells with benzamidine, a trypsin inhibitor, completely abolished any e ~ e activation. Adding a second enzyme substrate specific for cathepsin B (rhodamine 110, bis(CBZ-L-phenylalanyl-L-arginine amide) dihydrochloride), using excitation and emission wavelengths of 488 and 530 nm, respectively, allowed simultaneous measurement of activation of both enzymes. Under these conditions, both hyperstimulation with CCK and calcium reuptake inhibition with thapsigargin induced activation of both enzymes. These were colocalized within the same discrete apical compartments, but the activation of trypsin preceded activation of cathepsin B. These results demonstrate the intracellular activation of digestive enzymes occurring within the granular region of the acinar cell early after application of a noxious stimulus, and show that trypsin is the first enzyme to be activated. Antioxidant Profiles in Chronic and Recurrent Acute Pancreatitis G Morris-Stiff, D J Bowrey, M Davies, G Clarke, D Oleesky1, M C A Puntis Departments of Surgery, University of Wales College of Medicine & Department of Biochemistry1, University Hospital of Wales. Evidence is emerging that the development of chronic pancreatitis may be related to a deficiency of natural antioxidants such as selenium, vitamin A and E. It is as yet unknown whether these same deficiencies are present in patients with recurrent acute pancreatitis or whether acute pancreatitis represents an intermediate between chronic pancreatitis and normal individ-

Volume 27, 2000

264

Abstracts CON (n = 19)

Selenium (µmol/L) Zinc (µmol/L) Copper (µmol/L) Vitamin A (µmol/L) Vitamin E (µmol/L) Alpha carotene (µmol/L) Beta carotene (µmol/L) Xanthine (µmol/L) Beta cryptoxanthine Lycopene (µmol/L)

0.75 (0.72–0.84) 11.4 (10.7–12.2) 17.4 (15.6–21.2) 2.56 (2.23–3.06) 27.5 (23.4–39.4) 0.02 (0.01–0.04) 0.29 (0.15–0.59) 0.25 (0.15–0.53) 0.08 (0.06–0.15) 0.30 (0.18–0.61)

RAP (n = 11) 0.77 (0.43–0.92) 14.5 (12.5–15.5)** 15.4 (14.3–17.9)*** 2.15 (1.95–2.61) 28,9 (21.8–40.0) 0.02 (0.01–0.04) 0.25 (0.12–0.28) 0.21 (0.11–0.38) 0.10 (0.07–0.16) 0.25 (0.14–0.33)

CP (n = 27) 0.44 (0.32–0.68)* 11.4 (9.4–13.8) 18 (17–20.2) 1.48 (1.22–2.04)* 13.0 (9.6–19.4)* 0.02 (0.01–0.03) 0. 13 (0.07–0.25)* 0. 1 5 (0.09–0.25)**** 0.05 (0.03–0.08)* 0.08 (0.05–0.12)*

*p < 0.05 CP vs. CON and CP vs. RAP; **p < 0.05 RAP vs. CON and RAP vs. CP; ***p < 0.05 CP vs. RAP; ****p < 0.05 CP vs. CON.

uals. The aim of this study was to determine the antioxidant status of patients with both acute and chronic pancreatitis and compare these groups with a control population. The study population consisted of 27 patients with chronic pancreatitis (CP), 11 patients with recurrent nongallstone acute pancreatitis (RAP) and 18 age- and sex-matched controls (CON). Demographic details were collected together with details of disease etiology where known. Biochemical assessments included the trace metals selenium, copper, and zinc; vitamins A and E; and the carotenoid series. The results are summarized in the table above. For patients with chronic pancreatitis there was no difference in the antioxidant profiles between those with alcohol-related and idiopathic forms of the disease. This study has identified significant deficiencies in a number of antioxidants in patients with chronic pancreatitis that are not seen in individuals with recurrent acute pancreatitis or in healthy controls. The results would suggest that patients’ antioxidant status may be an important indicator of the pattern of pancreatic disease experienced, and may offer a window for therapeutic intervention.

Methods. This prospective study consisted of 71 individuals (23 patients with AP, 20 controls with non-pancreatitic acute abdomen and 28 age- and sex-matched healthy volunteers). Severity classification was based on the Atlanta criteria. Serum and u-irCAPAP-B were measured for all patients on admission by RIA. s-L-selectin was measured by ELISA. AP and acute abdomen subjects were also assessed for APACHE II scores and CRP levels. The Mann-Whitney U test and receiver operating characteristic curve (ROC-curve) were used to analyze the results statistically. Results. See table on facing page. Conclusion. s-irCAPAP and u-irCAPAP were significantly higher in severe AP patients. On admission, both were more sensitive and specific than APACHE II and CRP as severity markers in AP on admission. However, s-L selectin did not predict disease severity.

Carboxypeptidase B Activation Peptide (CAPAP-B) and L-Selectin as Severity Markers in Acute Pancreatitis

S Wemyss-Holden, S A White, C D Sutton, D P Berry, A R Dennison

Spleen Preserving Total Pancreatectomy For Chronic Pancreatitis

Department of General Surgery, Leicester General Hospital. C T Ung, S L Westlake, H J Brennan, C D Johnson University Surgical Unit, Southampton General Hospital. Introduction. Acute pancreatitis (AP) is a common surgical emergency of unpredictable course. To date, an optimal severity marker in AP is still lacking. Carboxypeptidase B (CAPAP-B) concentration correlates with the activation of trypsinogen, which is involved in the pathogenesis of AP. L-selectin is a product of leucocyte activation, therefore it may play a central role in the pathogenesis of acute inflammation. Aim. To assess irCAPAP-B in serum (s) and urine (u), and L-selectin in serum as severity markers in AP on admission. To compare the accuracy of these markers with APACHE II score and C-reactive protein (CRP) levels in predicting severe AP.

International Journal of Pancreatology

Aim. It is often considered impossible to preserve the spleen at the time of distal or total pancreatectomy when needed for chronic pancreatitis (CP). The aim of this study was to assess the feasibility of preserving the spleen in patients requiring total pancreatectomy for CP. Method. All patients requiring total pancreatectomy for CP were evaluated by a surgeon, gastroenterologist, and endocrinologist if simultaneous auotransplantation was also considered appropriate. Preoperative investigations included abdominal ultrasound, abdominal computed tomography, endoscopic retrograde cholangiopancreatography, and laparoscopy. Postoperatively, all patients underwent abdominal ultrasound and power Doppler to assess splenic perfusion and the patency of the remaining splenic vessels.

Volume 27, 2000

Oral Presentations

265 Median (interquartile range) on admission

s-irCAPAP* (nmol/L) u-irCAPAP* (nmol/L) s-L Selectin** (ng/mL) APACHE II+ CRP++ (mg/L)

Mild

Severe

Control

2.4 (0.3–6.5) 11.5 (2.4–68) (n = 15) 558.5

7.2 (6.3–17.3) 134.4 (35–46.9) (n = 7) 537.4 (301.5–800.3) 9 (5–11) (n = 7) 16.6 (12.2–292.5)

0.7 (0.5–2.5) 8.5 (3.1–50.2) (n = 3) 445.9 (435.3–735.3) 3 (0–13) (n= 1 4) 242.2 (77–250) (n = 8)

(294.4–999.5) 4 (0–11) (n = 16) 35.4 (6.4–832) (n = 13)

Volunteers 1.1 (0.6–1.4) 8.6 (3.7–11.1) (n=19) 1257 (575.3–1911)

Cut-off

Sensit %

Specific %

>6.0

100

86.7

>30

100

93.8

>5

71.4

75.0

>20

5.0

3.9

n = Number of patients. *p Value <0.05 for mild vs. severe, severe vs. controls, severe vs. volunteers. **p-value >0.05 for mild vs. severe, <0.05 for severe vs. controls, severe vs. volunteers +p value <0.05 for mild vs. severe, p-value >0.05 for severe vs. controls, ++p value >0.05 for mild vs. severe, severe vs. controls.

Results. Of 35 patients having total pancreatectomy, the spleen was preserved in 30 patients (19 F⬊11 M, mean age 40 yr). The etiology of CP was mainly idiopathic (n = 14) or alcohol (n = 12) related. All patients presented with chronic abdominal pain (mean 5 yr) and required opiate-derived analgesia for pain relief (e.g., morphine sulfate tablets, pethidine, dihydrocodeine). Fifteen patients (60%) had undergone previous therapeutic intervention for pain relief, drainage procedure (n = 5; stent, sphincterotomy or surgical), laparoscopic division of adhesions (n = 4), nerve blocks (n = 5) or distal pancreatectomy (n = 1). The spleen was preserved with an intact splenic artery and vein in 18 patients and in the remainder by the short gastric vessels (n = 11). In 1 patient the splenic artery was sacrificed, but the splenic vein remained intact. The mean duration of the procedure was 7 h (range 5–11 h) and mean blood loss was 1090 mL. The 30-d mortality was 4% (n = l). Two further patients died 6 mo postoperatively of unrelated causes. Five patients developed splenic complications (20%). These included splenectomy (n = 2), intrasplenic collection (n = 2) and a wedge splenic infarct. Two of these complications were related to intrasplenic islet autotransplants. At follow-up with splenic ultrasound and power Doppler, no other abnormalities could be detected and flow was demonstrable in all those with intact splenic arteries and vein. The mean duration of hospital stay was 25 d. Of the 24 patients alive beyond 6 mo follow-up, 82% (n = 20) have complete relief of pain, 4 still require opiate analgesia. Conclusion. Spleen-preserving total pancreatectomy is a feasible procedure and should be attempted in order to reduce the morbidity associated with an unecessary splenectomy. International Journal of Pancreatology

A Prospective Longitudinal Study to Assess the Reliability, Validity, and Responsiveness to Changes in Health State of the EORTC QLQ-C30 and QLQ-PAN26 in Patients with Pancreatic Cancer D Fitzsimmons, C D Johnson Pancreatic Biliary Research Unit, University of Southampton, UK Introduction. There is overwhelming consensus that quality of life (QoL) assessment is required in pancreatic cancer patients. A disease-specific QoL questionnaire, the QLQ-PAN26, has been developed to supplement the EORTC QLQ-C30 as a comprehensive method of assessment of QoL for international trials. Aim. To provisionally assess the reliability, validity and sensitivity of the QLQ-C30 and QLQ-PAN26 in patients with pancreatic cancer. Method. Patients were recruited soon after diagnosis, prior to chemotherapy or stent insertion (group A), or prior to surgical resection (group B). Patients were asked to complete QoL assessments at baseline, mo 1 and 3 or until withdrawal or death. Evaluation of the questionnaires was conducted using standard psychometric tests in SPSS. Results. Forty-five patients completed baseline QoL assessments, with 25 patients completing QoL assessments to mo 3. The main reason for dropout was disease progression and death in group A. Using known group comparisons, there was no significant difference in QoL scores between group A and B at baseline. At mo 1, group A showed a small improvement, whereas Volume 27, 2000

266

Abstracts

group B showed a decline in QoL scores. At mo 3, group A showed a decline in QoL, whereas group B showed an improvement towards pretreatment QoL scores, although statistical significance was not reached. The internal consistency and scale formation of the QLQ-C30 and QLQ-PAN26 was confirmed with Cronbach’s α of 0.6–0.9 across multi-item scales and factor analysis. Conclusion. The QLQ-C30 and QLQ-PAN26 is a reliable and valid method of QoL assessment in pancreatic cancer patients and can detect small changes in QoL over time, when stratified by treatment intention. This study highlights the difficulties with missing data in QoL assessment in this population of patients. A series of studies is progress to further evaluate the QoL questionnaires in an international population of patients.

Detection of Nodal Micrometastases in Pancreatic Cancer A Loverseed, *R M Charnley, M K Bennett Department of Histopathology and *Surgery, Freeman Hospital, Newcastle upon Tyne, UK Background. Survival after resection for pancreatic adenocarcinoma is very poor for node-positive patients. Long-term survival is mainly confined to a small proportion of nodenegative patients, but even in this group the cure rate is less than 25%. In order to improve our staging of patients following resection, we have evaluated the role of immunocytochemistry in detecting nodal micrometastases. Methods. From the pancreatic cancer database all NO tumors were selected. Nine cases were available. TI (2 cases), T2 (2 cases), and T3 (5 cases) tumors were studied. A median of 16 nodes per case could be recovered for analysis (range 1–25). Each lymph node specimen had 10 serial sections (5 µ thick) cut and stained routinely for haematoxylin and eosin (H&E) at levels 3, 6, and 9 and by immunocytochemistry for a mixture of cytokeratins at levels 4, 7, and 10. Results. Micrometastases were found in 4 of the 9 cases (all 4 were T3 tumors), either confined to the subcapsular sinus as individual cells or small aggregates of tumor, ranging in size from 27–642 µ (mean 196 µ). Routine histology (H&E) of these deeper sections demonstrated micrometastases, in 2 cases but immunocytochemistry found tumor emboli in all 4 cases. Conclusions. Improvements in the detection of small nodal deposits of tumor can be best achieved by immunocytochemistry using antibodies against a mixture of cytokeratins. The functional and prognostic significance of these deposits is at present unclear, but based on these results, 4 out of 5′T3 NO′ tumors should be reclassified as ′T3 NI′ tumors.

International Journal of Pancreatology

Expression Patterns of Neurotrophins and Trk Receptors in Human Pancreatic Cancer and Their Relationship to Perineural Invasion M Manu, D Mirza, J A C Buckels, and S R Bramhall Department of Hepatobiliary Surgery, University Hospital Birmingham, UK Background. Perineural invasion in pancreatic cancer is seen in 90% of tumors and this may be responsible for the high rate of local recurrence seen after resection. The neurotrophins (NTs) are a group of neurotrophic factors that have been demonstrated to play a role in the invasiveness, chemotactic behavior, and tumor cell survival of both neuronal and nonneuronal cancers. Methods. Four human pancreatic cancer cell lines (Suit2, Capan2, Pancl, and Bxpc3) were screened by Western blotting for expression of the NTs: nerve growth factor (NGF), neurotrophin 3 (NT-3), and their Trk receptor. The same cell lines were exposed to erogenous NGF and NT-3 to assess any mitogenic effect. The cell lines were also evaluated for the effects of NGF and NT-3 on invasion and chemotaxis using Matrigel invasion chambers. All cell lines were subjected to reverse transcriptase-polymerase chain reaction analysis for expression of Trk receptor subtypes. Results. All four cell lines were found to express the Trk receptor. None of the cell lines expressed NGF or NT-3. NGF but not NT-3 was found to be mitogenic for all the cell lines. Two of the four cell lines showed increased invasiveness through the matrigel membrane when incubated with NGF at a concentration of 50 ng/mL. Conclusion. These data suggest that NTs may play a role in pancreatic cancer mitogenesis and that their expression and the presence of their receptor are implicated in the perineural invasion. Specific Trk receptor antagonists may have a role in the prevention of local recurrence.

The Expression of Matrix Metalloproteinases (MMPs 1, 2, 3, 7, 8, 9, 12 and 14) and Their Inhibitors (TIMPs 1, 2 and 3) in Pancreatic Adenocarcinoma L E Jones, F Campbell, J P Neoptolemos Departments of Surgery and Pathology, Liverpool University, UK Introduction. Matrix metalloproteinases (MMPs) and their inhibitors are a group of enzymes involved in the breakdown and remodeling of extracellular matrix proteins. Overexpression of MMPs in cancer leads to tumor invasion and metastatic spread. Aim. To investigate in detail the expression of MMPs and TIMPs (tissue inhibitors of metalloproteinases) in normal pancreas and in pancreatic cancer (PC). Methods. Formalin-fixed, paraffin-embedded blocks of normal pancreas and ductal adenocarcinoma were obtained from

Volume 27, 2000

Oral Presentations 55 resections. Immunohistochemistry was performed for antibodies against MMPs 1, 2, 3, 7, 8, 9, 12, 14 and TIMPs 1, 2, and 3 (Dako Envision method©). Results. All MMPs were expressed to some degree in both normal (10 cases) and cancerous tissue (45 cases), with the exception of MMP9 being expressed only in PC. The intensity of expression was significantly increased in adenocarcinoma cells for MMPs 2, 7, 8, 9, and 12 in comparison to normal epithelium ( p = 0.04, p < 0.0001, p < 0.0002, and p = 0.0027, p = 0.006 and p = 0.01, respectively), although MMP3 stromal activity was reduced in cancer specimens ( p <0.0001). TIMP I was expressed by normal pancreatic islets and neuroendocrine cells but was only expressed by 3/45 tumors. One tumorexpresssing TIMPI was shown to have neuroendocrine differentiation. TIMP 2 expression was faint in normal pancreas ductal epithelium and was expressed only faintly in focal tumor cells in 18/45 cases. There was up regulation of TIMP2 in normal pancreas adjacent to the tumor with an increased stromelysin (MMP3) activity (p = 0.029). TIMP3 was expressed strongly in normal pancreatic acini and islets, although expression was faint or absent in the ductal epithelium. There was an increased intensity of TIMP3 epithelial expression in cancer with 43/45 cases staining positive ( p <0.0001). Conclusion. Overall MMP activity appears to be increased in PC. The low production of TIMP1 and TIMP2 might explain the poor survival in PC.

International Journal of Pancreatology

267 A Pivotal Randomized Controlled Study Comparing Marimastat Against Gemcitabine as First Line Therapy in Patients with Nonresectable Pancreatic Cancer 4

1 S Bramhall, 1J Buckels, 2A Rosemurgy, 3R Charnley, D Lloyd, 4W Steward, 5H Scarffe, 6C Johnson, 7S Falk, 8 S Rosewicz, 9L Curtis and Marimastat Pancreatic Cancer Study Group

1

Birmingham, 2Tampa, 3Newcastle upon Tyne, 4Leicester, 5 Manchester, 6Southampton, 7Bristol, 8Berlin, 9 British Biotech Pharmaceuticals Ltd.

Pancreatic cancers express high levels of matrix metalloproteinases (MMPs) that are important in tumor cell invasion and metastasis. Marimastat is an orally active, broad-spectrum MMP inhibitor. A phase II study of marimastat in patients with advanced pancreatic cancer identified a suitable dose range for randomized study and revealed an acceptable side effect profile. Patients with pathologically confirmed, nonresectable and chemo-naïve pancreatic cancer were randomized to receive marimastat at 5, 10, or 25 mg bid or gemcitabine 1000 mg/m2 intravenously weekly until disease progression or toxicity. The primary endpoint was survival; time to disease progression, patient benefit assessment score, and toxicity were secondary endpoints. A total of 414 patients were recruited. Each of the four treatment arms was matched for mean age, sex, weight, Karnofsky Performance Score, and stage of disease. There was no statistically significant difference in survival between any of the four treatment groups. Using the Cox proportional hazards model, however, survival in the 25 mg marimastat- and gemcitabinetreated groups appeared to be significantly better than in either the 5 mg or 10 mg marimastat-treated groups. There was no significant difference between the 25 mg marimastat and gemcitabine groups. There was no significant difference in FACT G quality of life scores and the overall incidence of treatment related adverse events between the four groups (14%, 10%, 15%, and 20% of patients, respectively). There appears to be a dose-dependent effect of marimastat treatment on survival in patients with advanced pancreatic cancer with no significant difference in survival in patients treated with 25 mg marimastat compared with gemcitabine. This study supports the use marimastat in patients with advanced pancreatic cancer.

Volume 27, 2000