EARLY TUMOR NECROSIS FACTOR ANTAGONIST TREATMENT PREVENTS PERIANAL FISTULA IN PEDIATRIC CROHN’S DISEASE

Abstract

Abstract BACKGROUND Perianal fistulizing complications (PFCs) develop amongst approximately 1/3 of children with Crohn’s disease (CD), are difficult to treat, and commonly recur. Recent retrospective studies suggested PFCs may be preventable with early therapy. We aimed to evaluate PFC prevention in a prospective cohort. Based on prior work, we hypothesized early tumor necrosis antagonist (anti-TNF) treatment would be associated with reduced risk of PFC development. METHODS The RISK cohort prospectively enrolled children (<18 yr) newly diagnosed with CD from 28 sites in North America (2008-2012). We obtained data from the Crohn’s & Colitis Foundation’s IBD Plexus program. PFCs were defined as perianal fistula and/or abscess. Patients with PFCs at or before enrollment were excluded. Non-penetrating perianal lesions such as skin tags and fissures were collectively considered “perianal lesions”. We evaluated treatment from enrollment until 1mo prior to PFC development or end of enrollment. Treatment was categorized into 3 mutually exclusive groups: 1) No immunosuppression; 2) Immunomodulators (IM, thiopurines and/or methotrexate) but no anti-TNF; and 3) Anti-TNF (regardless of IM). Steroids were considered separately. We performed nearest neighbor propensity score trio matching, matching patients across treatment categories on gender, age at diagnosis, growth delay, deep endoscopic ulcers, small bowel involvement, and colonic involvement. Using matched samples, we evaluated the predicted probability of PFC development first by treatment group, then by time to follow up of the medication initiation after studying into 6-time intervals to maximum 3yr. Logistic regression analyses were used to estimate the odds ratios (OR) for PFC development. RESULTS In total, 889 patients without PFC at enrollment in RISK were included. Overall, 41% were female, 70% were >10 yr old. The study included a moderately ill population, including 21% with growth delay, 43% with deep ulcers, and 70% with weighted pediatric Crohn’s disease activity index (wPCDAI) >30. After propensity score matching, 621 remained, including 207 in each treatment group. Propensity score matching was well balanced except more anti-TNF users (76%) had growth delay (26%) than IM (16%) or neither (65%; p=0.002). Regression revealed anti-TNF therapy was associated with 79% reduced odds, and perianal lesions 3-fold increased odds of PFC development. No other characteristic or treatment were significant (Table). Predicted probability of PFC development by each treatment group is depicted in the Figure. CONCLUSIONS In a prospective inception cohort of children with CD who had not yet developed PFCs, only early anti-TNF therapy use was associated with reduced risk of PFC development. These findings corroborate other studies, and indicate early treatment prevents or delays PFC development. Table. Adjusted odds ratio estimates of perianal fistula complications (after propensity matching). Figure. Predicted probability of developing perianal fistulizing complication by treatment group.