S450 Poster presentations collected prior to the injection using a commercially available vali- dated enzyme-linked immunosorbent assay (ELISA). Continuous variables were tested using the Mann–Whitney U-test. Receiver operating characteristic (ROC) curves were used to estimate the cut-off of the UTK trough level; for scoring purposes, we chose a cut-off that would maximise the sum of sensitivity and specificity. A value of p < 0.05 was considered statistically significant. Statistical analysis was performed using SPSS for Windows. Results: Forty-two patients were included (27 men and 15 fe- male), with an average age of 47.7 years. The average FC levels were 1358,5μg/ml at the beginning of the treatment. In the main- tenance period, we observed a biochemical response in 69% of patients, whose UTL were significantly higher than in non-respond- ers (2.25 µg/ml [IQR: 3.08] vs. 0..65 µg/ml [IQR: 1.95]. respect- ively; p = 0.037). Likewise. 38% of patients achieved biochemical remission (responders 2.125 µg/ml [IQR: 2.25] vs. non-responders 1.5 µg/ml [IQR: 3.26]; p = 0.476). The AUC for predicting biochem- ical response by means of UTL was 0.703 [CI 95%: 0.529–0.877] (p = 0.037), with sensitivity of 51%, specificity of 76% and a cut-off of 2.20 μg/ml. Conclusion: There is limited data on the association between UTL and patient outcome. Our study demonstrates the association be- tween improvement of different biomarkers and higher levels of UTL, with a significant statistic result for biochemical response and non-significant statistic result for biochemical remission, maybe due to a small sample, but with a positive trend. Further studies are ne- cessary in order to conclude that UTL is significant in the manage- ment and evolution of the disease and to determinate the optimal cut-off. P521 Infection of the anal canal by human papillomavirus in Crohn’s disease V. Guzela* 1 , C. Sobrado Júnior 1 , S. Nadal 2 , C.R. Nadal 2 , L.L. Villa 3 , A.P. Gonçalves 1 , L.R. Boarini 1 , S.C. Nahas 1 , I. Cecconello 1 1 Hospital das Clínicas da Faculdade de Medicina da Universidade de São Paulo, Gastroenterologia - Divisão de Coloproctologia, São Paulo, Brazil, 2 Instituto de Infectologia Emílio Ribas, Departamento de Coloproctologia, São Paulo, Brazil, 3 Faculdade de Medicina da Universidade de São Paulo, Laboratório de Inovação em Câncer do Centro de Investigação Translacional em Oncologia, São Paulo, Brazil Background: Compare the frequency of subclinical papillomavirus (HPV) in the anal canal os Crohn´s disease (DC) patients to a con- trol group by smear cytology and polymerase chain reaction (PCR) and hybridisation and to assess whether there is correlation with the presence of immunosuppression and anal manifestations of the disease, besides establishing the agreement between the two diag- nostic methods used. Methods: Two groups were selected, one with DC and others to be the control population. All the individuals were submitted to smear cytology (two brushes inserted in sequence in the anal canal) and a third one preserved for molecular analysis (PCR and hybridisation). The cytology was classified according to Bethesda criteria as normal, atypical squamous cells of undetermined significance (ASCUS), low-grade squamous intraepithelial lesion (LSIL), high-grade squa- mous intraepithelial lesion (HSIL). The molecular analysis was con- sidered with HPV present or absent and types 16 and high-risk HPV group identified. Results: During sixteen months, 104 patients underwent anal cy- tology collection for smear preparation and HPV molecular method detection. The control group consisted of 41 individuals and the DC of 63 patients (32 immunocompetent and 31 immunosuppressed). ASCUS and LSIL represented respectively 26.8% and 22.0% of the control patients and 28.6% and 39.7% of the Crohn´s patients, with no statistically significant difference between the groups. HPV was identified in 17.1% of patients in the control group and 27% of patients with CD by the molecular method, being predomin- antly high risk in both groups. In a subanalysis considering Crohn’s disease patients with and without immunosuppression and the con- trol group, there was no statistical difference between the frequen- cies detected by PCR and hybridisation (p = 0.084) as well as the anal manifestation of Crohn’s disease. The Kappa coefficient for agreement between smear anal cytology and HPV identification by PCR and hybridisation was -0.127 in the total sample. Conclusion: In this study, is possible to conclude that in the sample studied there was no difference between subclinical HPV anal infec- tion between control patients and Crohn’s disease group evaluated by both cytology and hybridisation and immunosuppression and anal impairment did not influence these results either. Anal cytology and hybridisation did not show agreement represented by the Kappa coefficient in this population. P522 Prior pharmacotherapy patterns among patients with IBD in the USA initiating biologic therapy R. Bornheimer 1 , S. Hass 2 , A. NAG* 3 , G. Oster 1 1 Policy Analysis Inc., Health Economics and Outcomes Research, Brookline, USA, 2 H. E. Outcomes, Health Economics and Outcomes Research, Los Angeles, USA, 3 Shire- a Takeda company, Health Economics- Outcomes Research and Epidemiology, Lexington, USA Background: Treatment for inflammatory bowel disease (IBD: Crohn’s disease [CD] and ulcerative colitis [UC]) has shifted from symptom management using corticosteroids to targeted therapies such as tumour necrosis factor (TNF) inhibitors. An increase in therapeutic options warrants a better understanding of current treat- ment pathways. This study reports on the pharmacotherapy history of patients with CD and UC prior to initiation of an IBD-targeted biologic. Methods: Using the US IQVIA™ Real-World Data Adjudicated Claims Database, we identified adults (≥18 years) with CD or UC who initiated biologics indicated for these diseases from 1 January 2016 to 30 June 2017. Patients were stratified based on the most re- cent biologic received; the date of initiation was designated the index date. Patients with < 12 months of continuous medical coverage prior to the index date were excluded. Patterns of prior IBD-related pharmacotherapy, in terms of the number and type of biologics, and time spent receiving the biologic, were examined in relation to the index biologic for the study period 1 January 2011 to 30 June 2017. Results: In total, 3010 patients with UC and 5210 with CD were identified. Among patients with UC, 1285, 977, 587 and 161 ini- tiated treatment with adalimumab, infliximab, vedolizumab and golimumab, respectively. Among patients with CD, 2634, 1436, 737, 241 and 162 began treatment with adalimumab, infliximab, vedolizumab, certolizumab and ustekinumab, respectively. Most patients beginning treatment with infliximab or adalimumab were biologic-naïve (UC: 79.9%, 89.3%, respectively; CD: 77.8%, Downloaded from https://academic.oup.com/ecco-jcc/article-abstract/14/Supplement_1/S450/5705608 by guest on 24 July 2020