Ulcerative Colitis: Ho GT

Ho GT, Moodie FM, Satsangi J. · University Department of Medical Sciences, Western General Hospital, Edinburgh, UK. · Gut. · Pubmed #12692067 links to  free full text

Abstract: The interface between luminal contents and intestinal epithelium constitutes the largest area of interaction between the host and the environment. There is now strong evidence that the gene product of the multidrug resistant pump (MDR) plays a critical role in host-bacterial interactions in the gastrointestinal tract and maintenance of intestinal homeostasis. This review highlights the efflux mechanism in the intestinal epithelium which is mediated by the multidrug resistant pump, also known as P-glycoprotein 170. Current studies promise to provide further insights into the contribution of the MDR1 gene in the pathogenesis of inflammatory and malignant disorders of the gastrointestinal tract.

Ho GT, Lee HM, Brydon G, Ting T, Hare N, Drummond H, Shand AG, Bartolo DC, Wilson RG, Dunlop MG, Arnott ID, Satsangi J. · Western General Hospital, University of Edinburgh, UK. · Am J Gastroenterol. · Pubmed #19262524 No free full text.

Abstract: OBJECTIVES: Calprotectin is a granulocyte neutrophil-predominant cytosolic protein. Fecal concentrations are elevated in intestinal inflammation and may predict relapse in quiescent inflammatory bowel disease. We aim to investigate fecal calprotectin (FC) as a biomarker in predicting the clinical course of acute severe ulcerative colitis (ASUC). METHODS: In 90 patients with ASUC requiring intensive in-patient medical therapy (January 2005-September 2007), we investigated the discriminant ability of FC to predict colectomy and corticosteroid and infliximab nonresponse. All patients received parenteral corticosteroids as first-line treatment; 21 (23.3%) were also treated with infliximab (5 mg/kg), after failure of corticosteroid therapy. RESULTS: Of 90 patients, 31 (34.4%) required colectomy, including 11 (52.4%) of those treated with infliximab. Overall FC was high (1,020.0 microg/g interquartile range: 601.5-1,617.5). FC was significantly higher in patients requiring colectomy (1,200.0 vs. 887.0; P=0.04), with a trend toward significance when comparing corticosteroid nonresponders and responders (1,100.0 vs. 863.5; P=0.08), as well as between infliximab nonresponders and responders (1,795.0 vs. 920.5; P=0.06). Receiver-operator characteristic curve analysis yielded an area under the curve of 0.65 to predict colectomy (P=0.04), with a maximum likelihood ratio of 9.23, specificity 97.4%, and sensitivity 24.0% at a cutoff point of 1,922.5 microg/g. Kaplan-Meier analyses showed that using 1,922.5 microg/g over a median follow-up of 1.10 years, 87% of patients will need subsequent colectomy. CONCLUSIONS: This is the first data set to demonstrate that FC levels are dramatically elevated in severe UC. These data raise the possibility that this biomarker can predict response to first or second-line medical therapy in this setting.

Lees CW, Ali AI, Thompson AI, Ho GT, Forsythe RO, Marquez L, Cochrane CJ, Aitken S, Fennell J, Rogers P, Shand AG, Penman ID, Palmer KR, Wilson DC, Arnott ID, Satsangi J. · Gastrointestinal Unit, Western General Hospital, Edinburgh, UK. · Aliment Pharmacol Ther. · Pubmed #19132970 No free full text.

Abstract: BACKGROUND: Anti-TNF agents are now widely used in Crohn's disease (CD), and in ulcerative colitis (UC). AIM: To review the safety profile of anti-TNF agents in all patients treated with infliximab in Edinburgh from 1999 to 2007. METHODS: Complete data were available on 202/207 patients comprising 157 CD, 42 UC and three coeliac disease. Median follow-up was 2.4 years (1.0-4.9) with a total of 620 patient-years follow-up. About 19.1% of CD patients were subsequently treated with adalimumab. RESULTS: Seven deaths (3.3%) occurred in follow-up; only one death was <1 year post-infliximab (at day 72, from lung cancer). A total of six malignancies (three haematological, three bronchogenic) and six cases of suspected demyelination (three with confirmed neurological disease) were reported. In the 90 days following infliximab, 95 adverse events (36 serious) occurred in 58/202 (28.7%) patients. In all, 42/202 (20.8%) had an infectious event (22 serious) and 27/202 (13.4%) of patients had an infusion reaction: 19 acute (four serious) and eight delayed (three serious). CONCLUSIONS: Serious infections, malignancies and neurological disease complicate anti-TNF use in clinical practice. Although evidence for causality is unclear, potential mechanisms and predisposing factors need to be explored. In individual patients, the risk/benefit analysis needs to be carefully assessed and discussed prior to commencement of therapy.

Lees CW, Zacharias WJ, Tremelling M, Noble CL, Nimmo ER, Tenesa A, Cornelius J, Torkvist L, Kao J, Farrington S, Drummond HE, Ho GT, Arnott ID, Appelman HD, Diehl L, Campbell H, Dunlop MG, Parkes M, Howie SE, Gumucio DL, Satsangi J. · Gastrointestinal Unit, Molecular Medicine Centre, University of Edinburgh, Edinburgh, United Kingdom. · PLoS Med. · Pubmed #19071955 links to  free full text

Abstract: BACKGROUND: Ulcerative colitis (UC) and Crohn's disease (CD) are polygenic chronic inflammatory bowel diseases (IBD) of high prevalence that are associated with considerable morbidity. The hedgehog (HH) signalling pathway, which includes the transcription factor glioma-associated oncogene homolog 1 (GLI1), plays vital roles in gastrointestinal tract development, homeostasis, and malignancy. We identified a germline variation in GLI1 (within the IBD2 linkage region, 12q13) in patients with IBD. Since this IBD-associated variant encodes a GLI1 protein with reduced function and our expression studies demonstrated down-regulation of the HH response in IBD, we tested whether mice with reduced Gli1 activity demonstrate increased susceptibility to chemically induced colitis. METHODS AND FINDINGS: Using a gene-wide haplotype-tagging approach, germline GLI1 variation was examined in three independent populations of IBD patients and healthy controls from Northern Europe (Scotland, England, and Sweden) totalling over 5,000 individuals. On log-likelihood analysis, GLI1 was associated with IBD, predominantly UC, in Scotland and England (p < 0.0001). A nonsynonymous SNP (rs2228226C-->G), in exon 12 of GLI1 (Q1100E) was strongly implicated, with pooled odds ratio of 1.194 (confidence interval = 1.09-1.31, p = 0.0002). GLI1 variants were tested in vitro for transcriptional activity in luciferase assays. Q1100E falls within a conserved motif near the C terminus of GLI1; the variant GLI protein exhibited reduced transactivation function in vitro. In complementary expression studies, we noted the colonic HH response, including GLI1, patched (PTCH), and hedgehog-interacting protein (HHIP), to be down-regulated in patients with UC. Finally, Gli1(+/lacZ) mice were tested for susceptibility to dextran sodium sulphate (DSS)-induced colitis. Clinical response, histology, and expression of inflammatory cytokines and chemokines were recorded. Gli1(+/lacZ) mice rapidly developed severe intestinal inflammation, with considerable morbidity and mortality compared with wild type. Local myeloid cells were shown to be direct targets of HH signals and cytokine expression studies revealed robust up-regulation of IL-12, IL-17, and IL-23 in this model. CONCLUSIONS: HH signalling through GLI1 is required for appropriate modulation of the intestinal response to acute inflammatory challenge. Reduced GLI1 function predisposes to a heightened myeloid response to inflammatory stimuli, potentially leading to IBD.

Noble CL, Abbas AR, Cornelius J, Lees CW, Ho GT, Toy K, Modrusan Z, Pal N, Zhong F, Chalasani S, Clark H, Arnott ID, Penman ID, Satsangi J, Diehl L. · Gastrointestinal Unit, Molecular Medicine Centre, Western General Hospital, Edinburgh, EH4 2XU, UK. · Gut. · Pubmed #18523026 No free full text.

Abstract: OBJECTIVE: To investigate differential intestinal gene expression in patients with ulcerative colitis and in controls. DESIGN: Genome-wide expression study (41,058 expression sequence tags, 215 biopsies). SETTING: Western General Hospital, Edinburgh, UK, and Genentech, San Francisco, USA. PATIENTS: 67 patients with ulcerative colitis and 31 control subjects (23 normal subjects and 8 patients with inflamed non-inflammatory bowel disease biopsies). INTERVENTIONS: Paired endoscopic biopsies were taken from 5 specific anatomical locations for RNA extraction and histology. The Agilent microarray platform was used and confirmation of results was undertaken by real time polymerase chain reaction and immunohistochemistry. RESULTS: In healthy control biopsies, cluster analysis showed differences in gene expression between the right and left colon. (chi(2) = 25.1, p<0.0001). Developmental genes, homeobox protein A13 (HOXA13), (p = 2.3x10(-16)), HOXB13 (p<1x10(-45)), glioma-associated oncogene 1 (GLI1) (p = 4.0x10(-24)), and GLI3 (p = 2.1x10(-28)) primarily drove this separation. When all ulcerative colitis biopsies and control biopsies were compared, 143 sequences had a fold change of >1.5 in the ulcerative colitis biopsies (0.01>p>10(-45)) and 54 sequences had a fold change of <-1.5 (0.01>p>10(-20)). Differentially upregulated genes in ulcerative colitis included serum amyloid A1 (SAA1) (p<10(-45)) the alpha defensins 5 and 6 (DEFA5 and 6) (p = 0.00003 and p = 6.95x10(-7), respectively), matrix metalloproteinase 3 (MMP3) (p = 5.6x10(-10)) and MMP7 (p = 2.3x10(-7)). Increased DEFA5 and 6 expression was further characterised to Paneth cell metaplasia by immunohistochemistry and in situ hybridisation. Sub-analysis of the inflammatory bowel disease 2 (IBD2) and IBD5 loci, and the ATP-binding cassette (ABC) transporter genes revealed a number of differentially regulated genes in the ulcerative colitis biopsies. CONCLUSIONS: Key findings are the expression gradient in the healthy adult colon and the involvement of novel gene families, as well as established candidate genes in the pathogenesis of ulcerative colitis.

Lees CW, Heys D, Ho GT, Noble CL, Shand AG, Mowat C, Boulton-Jones R, Williams A, Church N, Satsangi J, Arnott ID, Anonymous00089. · Gastrointestinal Unit, Western General Hospital, Edinburgh, UK. · Aliment Pharmacol Ther. · Pubmed #17635376 No free full text.

Abstract: BACKGROUND: Forty per cent of patients with acute severe ulcerative colitis will not respond to intravenous corticosteroids and require second-line medical therapy or colectomy. A recent controlled trial has suggested that infliximab may be effective as rescue therapy. AIM: To assess the value of infliximab as rescue therapy for acute severe colitis in a retrospective cohort of ulcerative colitis patients in Scotland. METHODS: All patients satisfied Truelove and Witts criteria on admission, failed to respond to intravenous corticosteroids and received infliximab (5 mg/kg) as rescue therapy. Response was defined as need for colectomy at hospital discharge and by 90 days. RESULTS: A total of 39 patients (median age 31.7 years) were treated. 26/39 (66%) responded, avoiding colectomy during the acute admission, and were followed up for a median of 203 days (Interquartile range = 135.5-328.5). Hypoalbuminaemia was a consistent predictor of non-response on univariate and multivariate analysis. At day 3 of intravenous steroids, 9/18 (50.0%) with serum albumin <34 g/L had urgent colectomy vs. 1/13 (7.7%) >or=34 g/L (P = 0.02, OR = 12.0, C.I. 1.28-112.7). Two serious adverse events occurred - one death due to Pseudomonas pneumonia, and one post-operative fungal septicaemia. CONCLUSIONS: Infliximab represents a moderately effective rescue therapy for patients with acute severe ulcerative colitis. Serious adverse events, including death, do occur and should be discussed with patients prior to therapy.

Brady RR, Collie MH, Ho GT, Bartolo DC, Wilson RG, Dunlop MG. · Academic Coloproctology, University of Edinburgh, Western General Hospital, Edinburgh, UK. · Colorectal Dis. · Pubmed #17302914 No free full text.

Abstract: OBJECTIVE: Controversy surrounds the optimal surgical management of the distal rectal remnant during colectomy for ulcerative colitis (UC) and the potential benefit from the placement of a rectal catheter for remnant drainage. This study reviews the clinical outcomes of patients who have undergone colectomy for UC with intra-peritoneal closure of the rectal remnant. METHOD: Analysis of prospective data lodged on Lothian Surgical Audit databases from patients treated in a tertiary coloproctology unit over 11 years. RESULTS: One hundred and fifty-nine patients were identified, the mean age was 41.9 years, 63% were men. Failure of maximal medical therapy necessitated surgery for 78.1% patients, while 12.6% had acute perforation and 11.9% had toxic megacolon. Complications included five (3.1%) stump dehiscences, eight (5.0%) intra-abdominal/pelvic collections, four (2.5%) significant wound infections, three (1.9%) small bowel obstructions and three (1.9%) deaths. Within the follow-up period, 62.3% patients had an ileo-pouch anal anastomosis (IPAA), 7.5% patients had a completion proctectomy, 10.1% patients within the series had a retained rectal remnant after 1 year follow up, the remaining patients had less than 1 year follow up. CONCLUSION: The intra-peritoneal rectal stump following colectomy for UC is associated with low rates of pelvic sepsis and a high proportion of patients successfully proceeding to IPAA.

Ho GT, Chiam P, Drummond H, Loane J, Arnott ID, Satsangi J. · Gastrointestinal Unit, Western General Hospital, Edinburgh, UK. · Aliment Pharmacol Ther. · Pubmed #16842459 No free full text.

Abstract: BACKGROUND: Corticosteroids remain the mainstay of first-line therapy in active inflammatory bowel disease. AIMS: To determine the clinical outcome after the first corticosteroid-therapy and to identify factors which predict response/failure. METHODS: 216 (136 ulcerative colitis and 80 Crohn's disease) patients were identified in this 5-year inception cohort. The outcomes of early (30 days) and late (1 year) responses were used. Multivariate analyses were performed to identify factors associated with outcome. RESULTS: 86 (63%) and 60 (75%) ulcerative colitis and Crohn's disease required corticosteroid therapy, respectively. In ulcerative colitis, at 30 days, 69 (51%), 42 (31%) and 25 (18%) patients demonstrated complete response, partial response and no response, respectively. For Crohn's disease, these outcomes were observed in 32 (40%), 28 (35%) and 20 (25%). After 1 year, 75 (55%), 23 (17%) and 29 (21%) patients with ulcerative colitis demonstrated prolonged response, corticosteroid-dependence or required surgery, respectively. For Crohn's disease, these outcomes were observed in 30 (38%), 19 (24%) and 27 (35%) patients. Extensive ulcerative colitis was a predictor of surgery (P = 0.001, OR: 15.2). In Crohn's disease, inflammatory disease behaviour was negatively associated with surgery (P = 0.02, OR: 0.13). CONCLUSION: Although corticosteroids are effective, dependence/resistance remains common. Patients with extensive ulcerative colitis and fistulizing/stricturing Crohn's are most at risk of failing corticosteroid therapy.

Ho GT, Soranzo N, Nimmo ER, Tenesa A, Goldstein DB, Satsangi J. · Molecular Medicine Unit and Gastrointestinal Unit, University of Edinburgh, Western General Hospital, Edinburgh, UK. · Hum Mol Genet. · Pubmed #16434479 links to  free full text

Abstract: Several lines of evidence suggest a role for the multidrug resistance gene (ABCB1/MDR1) and its product, P-glycoprotein 170, in the pathogenesis of inflammatory bowel disease (IBD). In addition, P-glycoprotein activity determines bioavailability of many drugs used regularly in many medical specialties, and ABCB/MDR1 variation appears to be a critical pharmacogenetic determinant. We have utilized a gene-wide haplotype tagging approach to further define the identity of germ-line variations in the ABCB1/MDR1 gene contributing to IBD susceptibility. Six haplotype tagging single nucleotide polymorphisms (tSNPs) representing the haplotypic variations of the ABCB1/MDR1 gene were identified initially following the characterization of the haplotype structure of this gene in 24 Centre d'Etude du Polymorphisme Humain Caucasian trios. Genotyping was performed in 249 ulcerative colitis (UC) and 179 Crohn's disease (CD) patients and 260 healthy controls. Using log-likelihood analysis, we observed a highly significant association between the common haplotypes and UC (P=4.22 x 10(-7)) but not CD (P=0.22). This significant association was critically dependent on one tSNP, intronic variant rs3789243. All haplotypes with this variant retained a highly significant association (P=3.2 x 10(-7)-3.6 x 10(-12)), whereas significance was lost when rs3789243 was dropped in systematic haplotypic analysis. The effect of this tSNP was independent of C3435T SNP, previously suggested to be the critical variant in disease susceptibility and drug transport. The association with UC was shown to be strongest with the phenotype of extensive disease (P=1.7 x 10(-7)). This 'candidate gene' approach provides compelling evidence to support the contribution of the ABCB1/MDR1 gene in determining risk to UC but not to CD and provides new insights into the localization of the critical susceptibility determinants within the gene. In addition, these findings have potentially important implications in the application of pharmacogenetics across a range of common diseases, including HIV, epilepsy and colorectal cancer.

Noble CL, Nimmo ER, Drummond H, Ho GT, Tenesa A, Smith L, Anderson N, Arnott ID, Satsangi J. · Gastrointestinal Unit, Western General Hospital, University of Edinburgh, Edinburgh, Scotland. · Gastroenterology. · Pubmed #16344054 No free full text.

Abstract: BACKGROUND & AIMS: Recent data suggest that polymorphisms in the organic cation transporter (OCTN) genes OCTN1 (SLC22A4) and OCTN2 (SLC22A5) represent disease-causing mutations within the IBD5 locus (chromosome 5q31). We investigated associations with disease susceptibility, phenotype, and evidence for epistasis with CARD15 in 679 patients with Crohn's disease (CD) or ulcerative colitis (UC). METHODS: A total of 374 patients with CD, 305 patients with UC, and 294 healthy controls (HCs) were studied. Genotyping for single nucleotide polymorphisms IGR2096, IGR2198, and IGR2230, OCTN1 variant (SLC22A4 1672C-->T), and OCTN2 variant (SLC22A5 -207G-->C) was performed using the TaqMan system. RESULTS: The IBD5 OCTN1 and OCTN2 polymorphisms were in strong linkage disequilibrium (D', >0.959). IGR2198 variant allele frequency (49.1% vs 40.8%; P = .0046) and homozygosity (21% vs 14.8%; P = .044) were associated with CD versus HCs. Variant allelic frequency of OCTN1 (53.6% vs 43%; P = .0008) and OCTN2 (56.1% vs 48.4%; P = .0092) polymorphisms and homozygosity for the OCTN1/2-TC haplotype (28.4% vs 16%; P = .0042) were associated with CD versus HCs. IGR2198 homozygosity and TC homozygosity were associated with stricturing/penetrating disease at follow-up (P = .011 and P = .011, respectively) and disease progression (P = .038 and P = .049, respectively) on univariate analysis and with need for surgery on multivariate analysis (P = .016 and P = .004, respectively). In the absence of the IBD5 risk haplotype, no association of OCTN1/2 variants with CD was detected. No associations were seen with UC. CONCLUSIONS: The IBD5 locus influences susceptibility, progression, and need for surgery in CD. However, the contribution of OCTN1/2 variants is not independent of the IBD5 haplotype; a causative role for these genes remains plausible but is not yet proven. Further genetic, functional, and expression data are now required.

Ho GT, Seddon AJ, Therapondos G, Satsangi J, Hayes PC. · Scottish Liver Transplantation Unit, New Royal Infirmary, Little France, Edinburgh, UK. · Eur J Gastroenterol Hepatol. · Pubmed #16292093 No free full text.

Abstract: BACKGROUND AND AIMS: The course of ulcerative colitis (UC) following orthotopic liver transplantation (OLT) for primary sclerosing cholangitis (PSC) is unclear. We documented the nationwide experience of the course of UC, before and after OLT for PSC. METHODS AND RESULTS: A total of 470 liver transplants were performed for 413 patients between 1992 and 2003, in the Scottish Liver Transplantation Unit, UK. Twenty-six patients had co-existing UC/PSC. Of these, data from 20 patients were studied over a median period of 11.9 years before OLT and 4.4 years after OLT; of the others, four patients required colectomy prior to OLT, one died within 7 days of transplant, and one developed UC after transplant. A significantly higher relapse rate (number of relapses/year of follow-up) was seen after OLT (median 1.0 versus 0.3; interquartile range, 0.10-1.42 and 0.01-0.40, respectively; P = 0.007). The corticosteroids requirement (number of courses/year of follow-up) after OLT was also significantly higher (0.40 versus 0.10; interquartile range, 0.51-1.13 and 0.05-0.12, respectively; P = 0.003). Twenty per cent of patients (4/20) became corticosteroid dependent after OLT. Thirty-five per cent of patients (7/20) underwent colectomy after OLT: three for severe disease and four for neoplasia/dysplasia. Five patients (19%) developed neoplasia following OLT. CONCLUSION: Despite immunosuppression, UC follows a more aggressive clinical course after OLT and is associated with a high rate of neoplasia.

Ho GT, Nimmo ER, Tenesa A, Fennell J, Drummond H, Mowat C, Arnott ID, Satsangi J. · Gastrointestinal Unit, Western General Hospital, Edinburgh, Scotland, UK. · Gastroenterology. · Pubmed #15685540 No free full text.

Abstract: BACKGROUND AND AIMS: The MDR1 gene encodes P-glycoprotein 170, an efflux transporter that is highly expressed in intestinal epithelial cells. The MDR1 exonic single nucleotide polymorphisms (SNPs) C3435T and G2677T have been shown to correlate with activity/expression of P-glycoprotein 170. METHODS: This was a case-control analysis of MDR1 C3435T and G2677T SNPs in a large well-characterized Scottish white cohort (335 with ulcerative colitis [UC], 268 with Crohn's disease [CD], and 370 healthy controls). We conducted 2-locus haplotype and detailed univariate and multivariate genotypic-phenotypic analyses. RESULTS: The MDR1 3435 TT genotype (34.6% vs 26.5%; P = .04; odds ratio [OR], 1.60; 95% confidence interval [95% CI], 1.04-2.44) and T-allelic frequencies (58.2% vs 52.8%; P = .02; OR, 1.28; 95% CI, 1.03-1.58) were significantly higher in patients with UC compared with controls. No association was seen with CD. The association was strongest with extensive UC (TT genotype: 42.4% vs 26.5%; P = .003; OR, 2.64; 95% CI, 1.34-4.99; and T allele: 63.9% vs 52.8%; P = .009; OR, 1.70; 95% CI, 1.24-2.29), and this was also confirmed on multivariate analysis ( P = .007). The G2677T SNP was not associated with UC or CD. These 2 SNPs lie in linkage disequilibrium in our population (D', .8-.9; r 2 , .7-.8). Two-locus haplotypes showed both positive (3435T/G2677 haplotype: P = .03; OR, 1.44) and negative (C3435/2677T haplotype: P = .002; OR, .35) associations with UC. Homozygotes for the haplotype 3435T/G2677 were significantly increased in UC ( P = .017; OR, 8.88; 95% CI, 1.10-71.45). CONCLUSIONS: Allelic variations of the MDR1 gene determine disease extent as well as susceptibility to UC in the Scottish population. The present data strongly implicate the C3435T SNP, although the 2-locus haplotype data underline the need for further detailed haplotypic studies.

Ho GT, Mowat C, Goddard CJ, Fennell JM, Shah NB, Prescott RJ, Satsangi J. · University Department of Gastroenterology, Western General Hospital, Edinburgh EH4 2XU, Scotland, UK. · Aliment Pharmacol Ther. · Pubmed #15142197 links to  free full text

Abstract: BACKGROUND: The failure rate of medical therapy in severe ulcerative colitis is high. A risk index, to aid the identification of patients of not responding at an early stage to intravenous corticosteroid therapy, would be useful to facilitate second-line treatment or surgery. METHODS: We recruited 167 consecutive patients with severe ulcerative colitis between January 1995 and March 2002; and employed multiple logistic regression to analyse parameters within the first 3 days of medical therapy. We applied statistical modelling to formulate a risk score according to the likelihood of medical failure. RESULTS: Sixty-seven (40%) patients failed to respond to medical therapy. Multiple logistic regression analysis identified mean stool frequency and colonic dilatation within the first 3 days and hypoalbuminaemia as independent predictors of outcome (P < 0.001, 0.001 and 0.002 respectively). A numerical risk score was formulated based on these variables. Patients with scores of 0-1, 2-3 and > or =4 had a medical therapy failure rate of 11%, 43% and 85% respectively. Receiver-operator characteristic analysis of this score yielded area under curve of 0.88, with a sensitivity of 85% and specificity of 75% using score > or =4 in predicting non-response. CONCLUSION: This risk score allows the early identification of patients with severe ulcerative colitis who would be suitable for second-line medical therapy or surgery.