Article Text

Download PDFPDF

Inflammatory bowel disease free papers 049–062

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.


K. Bodger1, A.R. Dodson2, C. Tysk3, F. Campbell2, E. Lindberg3, G. Järnerot3, J.M. Rhodes1. Depts of1Medicine and2Pathology, University of Liverpool, UK;3Orebro Medical Centre Hospital, Orebro, Sweden

Introduction: Alterations in epithelial glycoprotein expression in inflammatory bowel disease (IBD) include increased expression of the oncofetal carbohydrate antigens, TF (galactose β1,3Nacetylgalactosamine) and sialyl-Tn (sTn, sialyl 2,6Nacetylgalactosamine). Previous chromatographic colonic mucin analysis in monozygotic twins with IBD suggested a possible genetic basis for these changes (Tysk et al, Gastroenterology 1991;100:419). The present study aimed to explore this further by assessing mucosal expression of specific glycoprotein epitopes amongst IBD twins.

Methods: Formalin-fixed rectal biopsies from 22 monozygotic twin pairs with IBD were studied (6 pairs concordant for UC or CD; 16 healthy twins). Affected twins were in clinical and endoscopic remission. Expression of sTn and TF was assessed by indirect immunohistochemistry using a specific monoclonal antibody and peanut lectin respectively, and compared with expression in rectal biopsies from normal mucosa (irritable bowel syndrome) controls (n=14 sTn; n=17 TF) with investigators blinded for diagnosis. TF positivity was scored from 0 to 3.

Results: Compared to histologically normal controls, unaffected twins showed significantly greater TF positivity and a trend towards greater sTn positivity (table).

Abstract 049

Of the 5 unaffected twins who were sTn positive, 4 had twins with UC and 1 with CD. One of the sTn-positive/TF positive healthy twins has subsequently developed UC.

Conclusions: The positive findings in unaffected twins support previous evidence of a biochemical mucin defect. This could be the result either of a direct genetically determined alteration in glycosylation or of a secondary, eg cytokine-mediated, alteration in glycosylation. The altered glycosylation could be relevant in determining changes in the mucosa-associated bacterial flora.


J.O. Lindsay1, C. van Montfrans2, A.A. te Velde2, F.M. Brennan1, H.J.F. Hodgson1, M.S. Rodriguez Pena2. 1Imperial College School of Medicine, London, UK; 2Academic Medical Centre, Amsterdam, Holland

Introduction: Recent therapeutic strategies for Crohn's disease (CD) have focused on modulating the immune response by targeting cytokines and their receptors. For example, daily injections of recombinant IL-10, a pivotal immunoregulatory cytokine, are safe and well tolerated, but have shown limited benefit in clinical trials. An alternative therapeutic approach using adenoviral vectors to deliver IL-10 has proven superiority over daily IL-10 injections in IL-10 -/- mice with colitis.

Aims: The aetiology of CD is more complex than the deletion of a single immunoregulatory gene. Thus, we investigated the therapeutic efficacy of adenoviral vectors encoding IL-10 (AdvmuIL-10) in mice with trinitrobenzene sulfonic acid (TNBS) induced colitis; a Th1 cell dependent disease that is also unresponsive to daily IL-10 injections.

Methods: Balb/C mice received a single i.v. injection of 1x108 PFU AdvmuIL-10, empty cassette virus (Adv0) or PBS. After 24 hours colitis was induced by two intra-rectal doses of 0.5 mg TNBS or vehicle separated by seven days. Mice were sacrificed 48 hours after the second dose of TNBS.

Results: Mice treated with AdvmuIL-10 suffered significantly less weight loss over time than Adv0 or PBS controls after TNBS administration (p<0.0001 by 2 way ANOVA). In addition, AdvmuIL-10 therapy led to a significant reduction in stool pro-inflammatory cytokine levels and acute phase response. Finally, the histological score of mice with TNBS colitis treated with AdvmuIL-10 was significantly lower than Adv0 or PBS treated controls (7.1±0.9 compared to 10.4±1.0 and 10.8±1.3 respectively; p<0.05). The therapeutic efficacy of AdvmuIL-10 was associated with a decrease in the TNF-α, IFN-γ and IL-6 levels detected in colonic homogenates from mice with TNBS colitis. AdvmuIL-10 had no effect on cytokine release from stimulated systemic lymphocytes.

Conclusion: AdvmuIL-10 offers superior therapeutic efficacy to daily IL-10 injections in the TNBS model of colitis. Gene therapy strategies using adenoviral vectors encoding IL-10 may prove to be a potent therapy for chronic inflammatory conditions such as Crohn's disease.


S. Campbell, K. Kingstone, S. Ghosh. Gastrointestinal Unit, Department of Medical Sciences, University of Edinburgh, Western General Hospital, Crewe Road, Edinburgh EH4 2XU, UK

Background: It is well recognized that patients with low TPMT activity are more susceptible in developing bone marrow suppression side effects. In the UK, it is not uncommon to prescribe AZA at doses lower than 2 mg/kg body weight, but the relationship of the various dosing regimens on effectiveness of maintenance with reference to TPMT activity has not been investigated. We aimed to find the impact of TPMT activity on the clinical course of IBD patients treated with low dose azathioprine (AZA, <2mg/kg).

Methods: We measured TPMT activity from blood samples from 113 IBD patients who were taking AZA, discontinued AZA because of side effects or had never taken AZA. TPMT activity was compared with 17 healthy controls. Relapse rates and time to first relapse were compared in IBD patients and stratified according to their TPMT activity.

Results: Patients who became neutropenic had a significantly lower mean TPMT activity than the mean TPMT activity of patients whom developed other side effects (ANOVA, p<0.05). Patients who became neutropenic within the first 4 months maintained this degree of neutropenia throughout AZA therapy. Survival curves were constructed (time to first relapse) for low-dose AZA treated patients for TPMT activity of <20 nmol/hour/ml and >20 nmol/hour/ml. There were a lower number of relapses in IBD patients with lower TPMT levels (p<0.05).

Conclusion: Mean TPMT activity was significantly lower in patients on a low dose of AZA in remission compared with those that relapsed. TPMT activity was significantly lower in patients who discontinued AZA due to neutropenia than those who discontinued due to other side-effects. Though speculative, it is possible that the higher dose of AZA is only necessary in patients with higher TPMT activity. Our study results also provide an explanation for the commonly observed phenomenon of prolonged remissions on a low dose of AZA in a proportion of UK IBD patients.


T. Ahmad1, G. Wallace2, T. James3, M. Bunce4, K. Mulcahy-Hawes1, A. Armuzzi1, F. Fortune5, M. Stanford2, K. Welsh6, S. Marshall4, D. Jewell1.1 Departments of Gastroenterology, 4Transplant Immunology, University of Oxford; 2Dept of Ophthalmology, St Thomas Hospital, London; 3Dept of Ophthalmology, Calderdale Royal Hospital, Halifax; 5Department of Oral Medicine, St James Hospital, Leeds; 6National Heart and Lung Institute, Imperial College, London, UK

Background: BD is a chronic multisystem inflammatory disorder associated with gastrointestinal tract inflammation in up to 50% of patients. Experimental and clinical evidence, most notably the efficacy of the anti-TNF agents infliximab, thalidomide and pentoxifylline, implicates TNF in disease pathogenesis. Association with HLA-B*51 has been reported worldwide but the relative risk varies widely suggesting this may be due to linkage disequilibrium (LD) with polymorphisms in nearby genes including TNFA.

Aims: To determine whether functional TNFA promoter polymorphisms are associated with susceptibility to BD.

Methods: LD mapping of 140 polymorphisms across 12 genes was carried out using PCR-SSP. Disease associations were evaluated at each biallelic SNP and by haplotype. 149 Caucasian BD patients and 350 healthy control subjects were studied.

Results: TNF-1031C was associated with disease (Pc=0.00003; RR 2.4; This allele was found on three TNF promoter haplotypes (T3, T6, T7) associated with disease (T3: P=0.03, RR 1.5, CI 1.1–2.3; T6: P=0.04, RR 3.5, CI 1.1–12.7; T7: P=0.001, RR 2.5, CI 1.4–4.3). Extended HLA haplotypes based on T3 and T7 were constructed. Peak RR on these was found at B*51 (RR 4.6 CI 2.8–7.7) and B*57 (RR 3.1 CI 1.7–5.6) respectively. The association with B*51 was found to be independent of TNF-1031C. Subgroup analysis of HLA-B*51, B*57 negative patients however revealed that the association with TNF-1031C remained (P=0.024; RR=1.9 CI 1.1–3.2).

Conclusions: (1) TNF-1031C is associated with susceptibility to Caucasian BD independently of the recognised association with B*51 and the novel association with B*57. (2) Transracial and functional studies are now required to dissect out the complex linkage disequilibrium demonstrated by this study.


T. Ahmad1, A. Armuzzi1, M. Bunce2, K. Mulcahy-Hawes1, S. Marshall2, T. Orchard1, J. Crawshaw1, O. Large1, A. de Silva1, J. Cook2, M. Barnardo2, S. Cullen1, K. Welsh3, D. Jewell1.1Departments of Gastroenterology, 2Transplant Immunology, University of Oxford, Oxford; 3National Heart and Lung Institute, Imperial College, London, UK

Background: Crohn's disease (CD) is characterised by extensive heterogeneity in terms of disease location, behaviour and response to treatment. The major focus of recent genetic research in CD has been the identification of susceptibility rather than phenotype determining genes. NOD2 on chromosome 16 and the HLA region on chromosome 6 have been associated with disease overall but there are no data regarding contribution to specific disease subtypes.

Methods: We studied 240 accurately characterised Caucasian CD patients who had been followed up at a single centre for a median time of 16 years and 354 healthy controls. Three NOD2 variants (Arg702Trp, Gly908Arg, Leu1007fsinsC) were studied and linkage disequilibrium mapping applied across 340 HLA polymorphisms, broken down into 24 discrete gene haplotypic blocks (alleles). Genetic comparisons were made between CD patients and healthy controls.

Results: The NOD2 variants were associated with ileal disease only (P<0.0001; RR=4.1). All 42 patients who possessed Leu1007fsinsC had ileal disease (P=0.0001). The risk of ileal disease was greatest in compound heterozygotes and homozygotes (P<0.0001; RR=37.4). Early age of onset was associated with carriage of Leu1007fsinsC (p=0.006) and compound heterozygosity (p=0.03). In contrast alleles on specific extended HLA haplotypes determine overall susceptibility (Cw*0802, P=0.0004, RR=3.05; DRB1*0701, P=0.03, RR=1.61), location (Perianal: MICA*010, P=0.01, RR=2.1; Colonic: BAT1A, P=0.0003, RR=3.6) and behaviour (Fistulating disease: DRB1*0103, P=0.02, RR=3.4).

Conclusions: (1) The clinical heterogeneity of CD may be defined by specific genotypes. (2) Patient stratification by such a molecular classification may lead to a better understanding of the different mechanisms that underlie this clinical heterogenetity.


T. Ahmad1, G. Wallace2, T. James3, M. Bunce4, K. Mulcahy-Hawes1, A. Armuzzi1, F. Fortune5, M. Stanford2, K. Welsh6, S. Marshall4, D. Jewell1.1Departments of Gastroenterology, 4Transplant Immunology, University of Oxford; 2Dept of Ophthalmology, St Thomas Hospital, London; 3Dept of Ophthalmology, Calderdale Royal Hospital, Halifax; 5Department of Oral Medicine, St James Hospital, Leeds; 6National Heart and Lung Institute, Imperial College, London, UK

Background: BD is a chronic multisystem inflammatory disorder that shares many clinical features with Crohn's disease (CD), including gastrointestinal inflammation, oral ulceration, uveitis and erythema nodosum. There are however important differences in prognosis and treatment. Association with HLA-B*51 and susceptibility to BD has been reported worldwide although in Caucasian BD this allele is present in only one third of patients. Three groups have reported associations between variants in the NOD2 gene and susceptibility to CD.

Aims: To determine whether NOD2 LRR variants are associated with susceptibility to Caucasian BD and to determine whether HLA-B*51 and NOD2 typing may be used to distinguish BD from CD.

Methods: PCR-SSP was used to determine HLA-B types and NOD2 variants (non-synonymous SNPs - Arg702Trp and Gly908Arg; frameshift mutation - Leu1007fsinsC). 129 Caucasian BD patients, 244 CD patients and 350 healthy Oxfordshire control subjects were studied.

Results: HLA-B*51 was significantly associated with BD only; 32.4% vs. 9.9%; p=2x10–10. No association was found between the NOD2 variants and susceptibility to BD despite stratification by B*51 status. When BD and CD patients were compared possession of a NOD2 variant without HLA-B*51 carried a likelihood ratio (LR) of CD rather than BD of 6.8. Conversely possession of HLA-B*51 without a NOD2 variant carried a LR of 4.7.

Conclusion: (1) Polymorphisms in the LRR of the NOD2 gene, associated with CD, do not confer susceptibility to Caucasian BD. (2) NOD2 and HLA-B*51 typing may be used to molecularly distinguish BD from CD. (3) Polymorphisms in other regions of the NOD2 gene need to be studied in BD before NOD2 can be excluded as a susceptibility gene for BD.


D.A. van Heel, D.P. McGovern, N.J. Lench, D.P. Jewell.Wellcome Trust Centre for Human Genetics, University of Oxford, UK

Introduction: NOD2 has recently been identified as the susceptibility gene for Crohn's disease on chromosome 16. Three distinct NOD2 mutations (Arg702Trp, Gly908Arg and Leu1007fsinsC) have been reported to be independently associated with Crohn's disease in European Continental and North American populations. NOD2 is a monocyte protein, which recognises intracellular bacterial components and through NF-kB activation stimulates cytokine production.

Aims: To assess the relative importance of the three NOD2 mutations, previously associated with Crohn's disease (CD), in the UK population.

Methods: We genotyped 587 IBD families (containing 252 UC, 294 CD trios, 321 unrelated CD cases) and 239 healthy controls (HC) (all UK Caucasian). Family based (ASPEX transmission disequilibrium test, TDT) and case-control association analyses were performed (unrelated cases were selected one per family, at random, and an average taken of 1000 such selections). Population attributable risk (PAR) was calculated.

Results: Significant associations with Crohn's disease were seen for the Arg702Trp (TDT Transmitted/Untransmitted 65/37, P=0.01, allele frequency in unrelated CD cases 10.6% vs. HC 2.8%) and Leu1007fsinsC (47/9, P<0.0001, 7.1% vs. 2.2%) mutations. No association was seen with Gly908Arg (11/12, P=0.1, 1.9% vs. 0.7%). Estimates of PAR were Arg702Trp 14%, Leu1007fsinsC 10%, all tested mutations 23%. Genotype relative risks of Crohn's disease were: Arg702Trp heterozygotes 3.5, homozygotes 58.9; Leu1007fsinsC heterozygotes 2.7, homozygotes 53.7; possession of any 2 mutant alleles 38.2.

Conclusion: The Arg702Trp variant is the most common NOD2 mutation in UK Crohn's disease, and carries the greatest relative risk and population attributable risk. The biological function of the Arg702Trp mutation is unknown. The Leu1007fsinsC mutation is also common in UK Crohn's disease, whereas the Gly908Arg mutation is extremely rare. The PAR calculation suggests there would be 23% less Crohn's disease in the UK if these NOD2 mutations had not arisen.


A. Armuzzi1, T. Ahmad1, S. Marshall 2, K. Mulcahy-Hawes1, O. Large1, J. Crawshaw1, K.L. Ling1, A.P. de Silva1, M. Bunce2, K. Welsh3, D. Jewell1.1 Departments of Gastroenterology, 2Transplant Immunology, University of Oxford; 3National Heart and Lung Institute, Imperial College, London, UK

Background: The HSP-70 gene family comprises three genes (HSP70-1, HSP70-2 and HSP70-Hom) located in the HLA class III region, an area implicated in determining disease susceptibility and phenotype in both Crohn's disease (CD) and ulcerative colitis (UC). The encoded proteins, expressed in response to cellular stress, are involved in intracellular protein folding and the chaperoning of peptides through the endoplasmic reticulum.

Aims: To determine whether haplotypes constructed from SNPs in two HSP70 genes are associated with disease susceptibility, location and behaviour of IBD.

Methods: We studied 580 accurately characterised Caucasian IBD patients (263 CD, 317 UC) and 341 healthy controls. Three synonymous HSP70-1 SNPs (A-110C, G190C, C438T) and two non-synonymous HSP70-Hom SNPs (T2437C, G2763A) were studied using PCR-SSP.

Results: Six HSP70 gene haplotypes were constructed from the five biallelic SNPs. H1 (AGTCG) conferred protection to CD overall (p=0.04, RR 0.7, CI 0.5-0.9). H5 (CCCTG) was associated with colonic disease (p=0.01, RR 1.62, CI 1.1–2.3) and H2 (AGCCG) with ileal disease (p=0.03, RR 1.8, CI 1.05–3.2). A negative association with H1 and perianal disease (p=0.008, RR 0.53, CI 0.3-0.8) was also found. When disease behaviour was analysed, H4 (AGCTG) (p=0.04, RR 1.51, CI 1.02–2.1) was associated and H1 (p=0.04, RR 0.65, CI 0.4–1.0) protective for stenotic ileal disease. No association was found with HSP70 gene haplotypes and UC susceptibility or phenotype.

Conclusion: Specific HSP70 gene haplotypes are associated with clinical CD phenotypes. Functional studies are required to clarify whether these associations are due to linkage disequilibrium with SNPs in other HLA genes.


K.L. Ling, C.J. Ooi, W. Luman, W.K. Cheong, H.S. Ng (introduced by D.P. Jewell).Department of Gastroenterology, Singapore General Hospital, Outram Road, Singapore

Introduction: The clinical features and natural course of ulcerative colitis (UC) in older patients is controversial. Reports in Western literature before 1980 suggested that UC in older patients is more aggressive, but recent studies suggest the opposite. There have been no reports on Orientals.

Aim: Describe and compare the features and disease course of Chinese UC patients in Singapore diagnosed before and after 50 years of age.

Methods: The notes of all Chinese UC patients followed up in the Singapore General Hospital for the last 30 years were reviewed. Late onset colitis was defined as UC diagnosed at or after 50 years of age. Extensive colitis was defined as UC extending proximal to the splenic flexure. Severity was determined using Truelove and Witts criteria.

Results: One hundred and thirty eight patients were diagnosed before 50 years of age (92 men 46 women, mean age 32), and 37 diagnosed at or after 50 (19 men 18 women, mean age 60). They were followed up for at least 1 year. Mean time to diagnosis from symptom onset was 8 months for younger patients, and 14 months for older patients. Significantly more older patients had extensive disease at diagnosis, 18(50%) versus 36(26%) in younger patients (p=0.019). Thirteen (9%) younger patients had severe disease at diagnosis, compared with 9(24%) older patients (p=0.023). Nine (25%) patients with late onset colitis only had the one initial attack of colitis versus 51(37%) of younger patients. A greater proportion of older patients (22%) were steroid dependant or needed a steroid sparing agent compared with younger patients (12%). These did not reach statistical significance. Nine (24.1%) patients with late-onset colitis required proctocolectomy for fulminant colitis or poorly controlled disease versus 15 (10.9%) patients with early-onset disease (p=0.034).

Conclusions: Chinese patients with late onset UC have significantly more severe and extensive disease at onset. They are more likely to require surgery for fulminant colitis or poorly controlled colitis compared with younger patients.


D. Joy, S. Ghosh.GI Unit, Western General Hospital, University of Edinburgh, UK

Introduction: The incidence of juvenile onset Crohn's disease (CD) continues to rise in Scotland over the past 3 decades. A recent Danish study has reported that month of birth may be linked to risk of CD later in childhood.1 This would appear to incriminate potential infectious agents operating in utero or early in childhood.

Objective: To investigate the influence of birth date on the development of CD later in life.

Methods: From the Scottish Hospitals Discharges Linked Database we identified 438 patients with juvenile onset CD (age of symptom onset before or at 16 years of age) between 1981 and 1995. This covered the entire Scottish population. All case notes were retrieved and the diagnosis verified. The month of birth distribution of the CD cases was compared to that expected from population statistics obtained from the General Register Office for Scotland. Seasonal trends were analysed using the von Mises distribution.

Results: The table shows the seasonal pattern by month of birth. The value within brackets gives the expected number of CD births. 52% were born in the first half of the year; there were 226 observed CD births between January and June against expected number of 220 (obs/exp = 1.03). 48% were born in the second half of the year; there 212 observed CD births between July and December against expected number of 219 (obs/exp 0.96). 26% were born in winter Dec-Feb), 23% in spring (Mar-May), 27% in summer (Jun-Aug) and 24% in autumn (Sept-Nov). Seasonal analysis using the von Mises distribution gave a χ2 = 3.67, p = ns.

Abstract 058

Conclusion: Our data, inclusive of the entire Scottish population aged 0–16 years, do not support significant influence of in utero or perinatal exposure to seasonal environmental/infectious agents in the later onset of juvenile CD.

1BMJ 2001;323:907.


A.B. Hawthorne1, S.P.L. Travis2, and BSG IBD Clinical Network.1University Hospital of Wales, Cardiff; 2John Radcliffe Hospital, Oxford, UK

Data on outcome of in-patient management of severe colitis is mainly available from single-centre studies in specialist hospitals. In preparation for a national clinical trial, prospective centres were asked to record clinical details of all patients admitted for intensive treatment of ulcerative colitis (UC) between May 1 and July 31 2001.

Methods: Data forms were sent to 45 centres intending to participate in the trial, and data collected on UC extent, duration, severity, outcome of treatment, and for patients undergoing colectomy – reasons for surgery.

Results: 116 patients, 53% male (median age 42, range 17–100) from 29 centres were reported. All received steroids. In 32 (28%) it was the first attack, and 78 (68%) fulfilled Truelove and Witts' criteria for a severe attack at admission. 47 (41%) responded completely (stool freq. ≤3), 36 (31%) had a partial response to treatment (stools >3, or visible blood), and 33 (28%) had a colectomy during that admission. On admission, median stool freq. (10/day), pulse (90bpm), temp. (37.0oC) and haemoglobin (12.3g/dl) did not differ significantly between the three groups. Admission CRP (n=78, median 63mg/l) did not predict response, but admission ESR (n=44) was significantly higher in the colectomy group (median 65mm/hr) vs complete response (34mm/hr) or partial response (20mm/hr) groups (Kruskall-Wallis, p=0.005). Outcome was not influenced by disease extent, duration or other factors. 15 received ciclosporin, with 9 avoiding colectomy. Follow-up questionnaires were received for 26 colectomy patients, who had surgery after a median 10 days in hospital. Four (15% had toxic dilatation, and 1 (4%) had perforation. The commonest reason for surgery was failure to remit after 7 days (77%). Three had procto-colectomy, 23 had subtotal colectomy. Conclusions: A colectomy rate of 28% for severe colitis across the UK is similar to that reported from single centres. At admission, ESR was more predictive of outcome than CRP. Duration, previous attacks, extent or severity of symptoms on admission did not predict outcome. Recruitment of 116 patients in 3 months makes large trials on severe colitis potentially viable.


T.R. Card1, A. Sawczenko2 , B.K. Sandhu1, R.F.A. Logan2.1Dept of Public Health and Epidemiology, Nottingham, UK; 2Department of Gastroenterology, Bristol Children's Hospital, UK

Background: Pre-natal or early neonatal exposures have been postulated to alter the risk of IBD later in life. The risk of Crohn's disease has been reported to vary with season of birth in cases reported from large UK referral centres. We report birth data from the 1998/9 BPSU-BSGRU UK incidence survey of IBD in those aged less than 20 years (Lancet 357;093–4).

Methods: The number of cases born in each of the 12 calendar months was corrected for the effect of variation in birth rate of the population of England and Wales for the 21 years during which the cases were born. The resulting monthly `rate' (numbers of cases born per month for each 100,000 live-births over the 21 years) was modelled using periodic regression.

Results: There were 659 cases of newly diagnosed Crohn's Disease (CD) and 297 cases of Ulcerative Colitis (UC). The periodicity of their births (fig 1) appeared similar to that of the total population. The periodicity regression models showed some evidence for residual periodicity after correction for this. (R2 =0.29 for CD and 0.43 for UC). However a likelihood ratio test showed these models not to be a significantly better description of the data than a straight line (chi2(6df) = 4.11,P= 0.6613 for CD, chi2(6df)=6.83, P=0.3371 for UC).

Conclusion: In this study, the largest population survey of childhood and adolescent IBD to date, we could not any detect significant periodicity in month of birth of either CD or UC cases.


H. Hanai1, H. Arai1, T. Furuta1, K. Yoshida1, A. Saniabadi2, F. Watanabe3, I. Matsushita4, K. Takeuchi5, T. Iida5 (introduced by B. Danesh).1Dept of Medicine, Hamamatsu University; 2Japan Immunoresearch Laboratories, Takasaki; 3Dept of Gastroenterology, Fujueda Munincipal General Hospital; 4Dept of Gastroenterology, Seirei General Hospital; 5Gastroenterology Centre, Hamamatsu South Hospital, Hamamatsu, Japan

Ulcerative colitis (UC) is an inflammatory bowel disease associated with activation of the immune system and inflammatory responses. Factors which initiate and perpetuate the inflammatory responses are not well understood. However, several recent reports have provided strong evidence for a major role by granulocytes and monocytes in the mucosal inflammation and UC relapse. Accordingly, we thought that patients with corticosteroid unresponsive severe UC might respond to granulocyte and monocyte apheresis (GMCAP). For GMCAP, we used a column of 335 mL capacity filled with 220g cellulose diacetate beads of 2 mm in diameter as the column adsorptive carriers (Adacolumn). The carriers selectively adsorb granulocytes and monocytes. Thirty one patients with prednisolone unresponsive severe UC and 8 corticosteroid naïve patients with severe UC received 10 GMCAP treatment sessions, one session/week for 10 consecutive weeks. Duration of one GMCAP session was 60 minutes, flow rate 30mL/minute. The efficacy of GMCAP was assessed by measuring UC clinical activity index, UC disease activity index and Matts' classification index of endoscopic mucosal appearance at baseline, week 6 and week 12. Patients who improved were given 6-mercaptopurine (30mg/day) to maintain remission. At week 12, 80.6% of corticosteroid unresponsive and 87.5% of corticosteroid naïve patients were in remission. Another 6.2% and 12.5% respectively, had their UC symptoms improved. Further, during the mean observation time of 8.8 months, no serious adverse side effects attributable to GMCAP were observed which is remarkably in contrast to cyclosporin A therapy for corticosteroid unresponsive UC. The major findings of this new treatment for UC are the followings, a) induced remission in patients with severe steroid refractory UC; b) reduced the number of patients who needed surgery; c) dramatically reduced the use of corticosteroid.


A.C. Ford, R.J. Towler, A.T.R. Axon, D.M. Chalmers.Centre for Digestive Diseases, Leeds General Infirmary, Leeds, UK

Background: Mycophenolate mofetil (MMF) has been claimed to be effective and well tolerated in refractory IBD although there is relatively little information regarding its use in clinical practice, particularly with reference to steroid sparing, toxicity, and longer term efficacy.

Aims: To review our experience in achieving and maintaining remission in refractory IBD and to document tolerability, major toxicity, and steroid sparing.

Methods: A retrospective audit was performed of the records of 20 patients treated with MMF over a 30-month period.

Results: Twenty patients (M=6 F=20, ages 18 to 72) were identified of whom 17 had Crohn's disease and 3 ulcerative colitis. All patients had been intolerant of, or had not responded to Azathioprine, and 19 were taking corticosteroids when MMF therapy was instituted. The median dose of MMF was 1.5g/day and mean duration of therapy was 10.7 months. MMF was discontinued in 12 patients –7 due to intolerance (4 non-specific symptoms, 1 joint aches, 1 lethargy, 1 skin rash) and 5 because of lack of efficacy. Of the 8 still on treatment at the end of the study period (mean duration of therapy 23.8 months) 6 were in remission (5 Crohn's, 1 UC) and off all steroid therapy, but 2 had relapsed and were being considered for alternative therapy. No major haematological, hepatic, renal, or GI toxicity was noted and there was no major sepsis. No predictors of response to MMF could be identified.

Conclusions: Approximately one third of patients with severe and refractory IBD achieved both remission and complete steroid withdrawal on MMF therapy. 35% of patients could not tolerate the drug, and a further third did not respond. No major toxicity was recorded. MMF therapy should be considered for patients refractory to steroids and Azathioprine, but longer term controlled studies are required.