P519 Higher anti-TNF α trough levels are not associated with increased radiological response in perianal fistulising Crohn’s disease: A multicentre study

T. LEE1, L. Gilbert2, A. Srinivasan3, A. Lee4, D. van Langenberg3, C. Martin5, N. Ding1, M. Sparrow2

1Gastroenterology, St Vincent’s Hospital, Melbourne, Australia, 2Gastroenterology, Alfred Health, Melbourne, Australia, 3Gastroenterology, Eastern Health, Melbourne, Australia, 4Eastern Health, Radiology, Melbourne, Australia, 5Monash University, Epidemiology and Public Health, Melbourne, Australia

Background

Perianal fistulas remain a debilitating and clinically challenging manifestation of Crohn’s disease (CD), given their prevalence and relative treatment resistance. Since the advent of biologic therapy, particularly the anti-TNFα agents infliximab and adalimumab, patient outcomes have improved. Serum trough levels have been associated with mucosal healing in luminal CD. However, the relationship between trough drug levels and healing of perianal fistulas remains less clear, with few studies assessing this cohort, and clinical healing typically defined as the primary endpoint. The aim of this study was to assess the relationship between radiological healing of perianal fistulising Crohn’s Disease (pfCD) on MRI, and serum trough drug levels, in patients treated with anti-TNFα therapy.

Methods

In this multi-centre, retrospective cross-sectional study, patients with pfCD who had trough levels measured within 6 months of a pelvic MRI were included. We collected patient demographics, infliximab and adalimumab trough levels, the presence or absence of anti-drug antibodies, concomitant steroid or antibiotic therapy, and Van Assche scores on MRI. The primary outcome, radiological response, was defined as a Van Assche score of 7 or less, while no response was defined as a score of greater than 7.

Results

A total of 99 patients were included (65 on infliximab, 34 on adalimumab). For patients receiving infliximab, the median drug levels for responders (n = 22) compared with non-responders (n = 43) were 5.5 mg/ml vs. 3.9 mg/ml respectively (p = 0.16). For patients receiving adalimumab, the median drug levels for responders (n = 14) compared with non-responders (n = 20) were 6.3 mg/ml vs. 3.1 mg/ml respectively (p = 0.09). On ROC curve analysis, the AUC for the association between radiological response and infliximab levels was 0.61, while for adalimumab it was 0.67. On quartile analysis, there appeared to be an increased response for increase in infliximab trough level quartile, with the exception of the highest quartile, however this association was not statistically significant (OR 1.36, 95% CI 0.85–2.16, p = 0.20). In comparison, the quartile analysis of adalimumab trough level and response demonstrated an exposure-response relationship (OR 2.24, 95% CI 1.11–4.52, p = 0.02).

Conclusion

Trough infliximab and adalimumab levels in patients who achieved radiological response were not significantly higher compared with those who did not, however quartiles analyses demonstrated trends toward an exposure-response relationship, in particular for adalimumab. The association between trough levels and radiological response could be further characterised with a larger, and ideally longitudinal, study.