allopurinol, (which blocks xanthine oxidase) appears to circumvent this problem, optimising both metabolite prole and clinical outcome. Data on the use of allopurinol and thiopurines remain limited. We previously reported our experience of using this combination in thiopurine-related hepatotoxicity. In this study we report further experience for a broader indication, including toxicity (mainly hepatic) non-response with adverse metabolite prole, and very high TPMT activity. Methods Patients prescribed combination treatment were identied from notes in the IBD clinic, TGN monitoring results and hospital pharmacy records. Data were collected retrospectively. We were particularly interested in whether combination treatment overcame the specic problem that precluded thiopurine monotherapy. We also analysed changes in metabolite proles and subsequent clinical outcome. Results 35 patients have been co-prescribed allopurinol and thio- purines since our previous report. 20 had Crohns disease and 15 ulcerative colitis, age range 20e65 years. Their average TPMT activity was 40 pmol/h/mgHb, (vs 32.5 in our population overall). In those receiving co-prescription for adverse effects (14 hepato- toxicity and three other: rash, nausea, fatigue) 94% were able to tolerate combination treatment with complete resolution of liver function tests abnormalities where relevant, of these 88% achieved clinical remission. In 10 patients failing thiopurine monotherapy 50% achieved clinical response on combination treatment. In eight started on combination treatment as primary therapy six (75%) achieved remission, comparing favourably with historical controls. The only adverse event was a transient lymphopaenia in a patient that had mistakenly continued full dose thiopurine. TGN levels increased from an average of 192 to 449 pmol/8310 8 RBC on co-treatment, while methylmercaptopurine levels decreased from an average of 503 pmol/8310 8 RBC on single agent treatment to just 16 on combination therapy. Conclusion Combination treatment with reduced-dose thiopurine and allopurinol circumvents predominant methylation, optimising metabolite proles, avoiding hepatotoxicity and improving clinical response. Using combination treatment as rst line should lead to a faster remission in those with high TPMT activity. Allopurinol co-prescription may also overcome other side effects. Whether combination therapy should be used rst-line in those with high TPMT activity, or more extensively, warrants further investigation in a prospective trial. Competing interests None. Liver symposium OC-016 SEQUENTIAL ACTIVATION OF CXCR3 AND CCR6 MEDIATES RECRUITMENT AND POSITIONING OF INTERLEUKIN 17 PRODUCING T CELLS IN THE LIVER doi:10.1136/gut.2009.208934p 1 Y H Oo, 1 V M Banz, 2 D Kavanagh, 3 D R Withers, 1 E Humphreys, 1 L Lee-Turner, 1 G Reynolds, 2 N Kalia, 4 S G Hubscher, 1 B Eksteen, 1 D H Adams. 1 Centre for Liver Research, Birmingham, UK; 2 Centre for Cardiovascular Sciences, Institute for Biomedical Research, Birmingham, UK; 3 School of Infection and Immunity, MRC Centre for Immune Regulation, Institute for Biomedical Research, Birmingham, UK; 4 Department of Path- ology, The Medical School, Institute for Biomedical Research, Birmingham, UK Introduction IL-17 secreting CD4 T cells (Th17) are a distinct lineage of pro-inammatory lymphocytes that develop under control of the transcription factor RORgt. Th17 cells are pathogenic in models of autoimmunity and have been implicated in the pathogenesis of several forms of liver inammation including primary biliary cirrhosis, alco- holic and viral hepatitis. Despite their importance in perpetuating inammation, there is little known about the molecular basis for their recruitment and positioning to specic sites such as the liver. Methods Human intrahepatic Th17 cells were isolated from explanted liver tissue without cytokine expansion and analysed by multi colour ow cytometry. Murine liver injury models of Con A hepatitis, MCD Diet and repeated CCL4 injections were used for intravital microscopic assessment of Th17 cell recruitment to the liver. Results Less that 0.25% of T cells in normal human liver expressed IL-17 but we detected IL-17 secreting cells in inamed human liver samples with the highest frequencies detected in patients with seronegative fulminant hepatitis (median 3.7%; IQR 3.4e3.9) non-alcoholic steatohepatitis (median 3.3%; IQR 2.7e3.5%) and auto-immune hepatitis (median 2.8%; IQR 2.6e3.3%). Intrahepatic IL-17+ cells expressed RORgt and the IL-23 receptor consistent with a Th17 lineage. They secreted IL-22 and IFNg in addition to IL-17 and expressed high levels of the chemokine receptors CCR5, CCR6, CXCR3 and CXCR6. Intravital microscopy in murine models of acute liver injury (con-A hepatitis), steatohepatitis (MCD) and chronic liver injury (CCl4) demonstrated enhanced recruitment of adoptively transferred Th17 cells to the hepatic parenchyma via the sinusoids. Recruitment was signicantly reduced by a blocking the CXCR3 ligand CXCL10 which also ameliorated liver injury. The CCR6 ligand, CCL20, was restricted to biliary epithelium and secreted by cholangiocytes in vitro suggesting a further recruitment signal positions Th17 cells near bile ducts in portal tracts. Conclusion Th17 cells are recruited to sites of inammation in the liver by sequential interactions in which the CXCR3 ligand, CXCL10 promotes recruitment into the parenchyma from blood via the sinusoids and the CCR6 ligand, CCL20 positions cells at the biliary epithelium. Competing interests None. OC-017 A COMPARISON OF NON-INVASIVE FIBROSIS SCORING SYSTEMS IN PATIENTS WITH NON-ALCOHOLIC FATTY LIVER DISEASE doi:10.1136/gut.2009.208934q 1 S McPherson, 1 S Stewart, 1 E Henderson, 2 A Burt, 1 C Day. 1 Liver Unit, Freeman Hospital, Newcastle upon Tyne, UK; 2 Department of Pathology, Freeman Hospital, Newcastle upon Tyne, UK Introduction Non-alcoholic fatty liver disease (NAFLD) affects approximately 30% of the population, with 3% having non- alcoholic steatohepatitis (NASH), which may progress to cirrhosis and liver failure. Accurate evaluation of liver brosis is important to identify patients who may develop complications. Liver biopsy is limited by its invasive nature. Therefore, several non-invasive strategies derived from simple clinical and laboratory parameters have been developed to identify patients with advanced brosis. The aim of this study was to compare the diagnostic performance of several non-invasive tests in patients with biopsy-proven NAFLD. Methods Patients who were reviewed in the Freeman Hospital liver clinic from 2003-2009 were included. Liver biopsies were assessed using the Kleiner score. The AST/ALT ratio, AST to platelet ratio index (APRI), BARD, FIB-4 and NAFLD brosis scores were calcu- lated from blood tests taken within 6 months of liver biopsy. Results 145 patients (82 male (61%), mean age 51612 years) were included. The mean BMI was 34.865 and 73 subjects (50%) had diabetes. 93 patients (64%) had NASH and 27 (19%) had advanced brosis (F3-4). The FIB-4 score had the best diagnostic accuracy for advanced brosis (AUROC 0.86), followed by AST/ALT ratio (AUROC 0.83), NAFLD brosis score (AUROC 0.81), BARD score (AUROC 0.77) and APRI (AUROC 0.67). Abstract 017 shows NPVs and PPVs for each test using specic cut-offs, along with the proportions of patients who could avoid liver biopsy by using each score. Gut April 2010 Vol 59 No 4 Suppl 1 A7 BSG abstracts on 4 May 2018 by guest. Protected by copyright. http://gut.bmj.com/ Gut: first published as 10.1136/gut.2009.208934p on 23 September 2015. Downloaded from