Effect of tumour necrosis factor antagonists on avoiding surgery in stricturing Crohn's disease: a tertiary centre real-life experience

Background Stenosis is the most common complication of Crohn’s disease (CD), often requiring surgical resection. Long-term outcome of patients receiving anti-TNF therapy for such disease complication is poorly known. Methods Enrolled were 51 CD patients followed-up in a tertiary IBD Centre between July 2006 and November 2015. All of them had stricturing CD (ileal = 49 and colonic = 2), diagnosed by colonoscopy and/or MRI enterography. Thirty-one subjects (61%) were given adalimumab and 20 (39%) infliximab. The primary outcome was the rate of surgery for stricturing CD. Statistical analysis included descriptive analysis, Wilcoxon test for differences in median values, logistic regression with univariate analysis for risk factors, and Kaplan–Meyer curves and Cox analysis for estimation of efficacy of anti-TNFs in avoiding surgical resection. All differences were considered statistically significant for p < 0.05. Results In the study, 51 CD patients (27 males; median age at diagnosis 31.9 yr [range 12–61]) were analysed. Median duration of disease at the beginning of anti-TNF therapy was 4.4 yr (range 0.3–37). The location of CD was ileal in 22 subjects, colonic in one, and ileocolonic in 28 patients. After a median follow-up period of 35.8 months (range 3–105), 20/51 patients (39%) underwent abdominal surgery. Amongst patients not undergoing surgery, half of them (25/51) continued anti-TNF therapy at last follow-up and 6/51 (11.7%) stopped anti-TNF treatment (1 for secondary loss of response, 1 for infusion reaction, 2 for psoriasis, 1 for pregnancy, and 1 for recurrent tonsillitis). Median values of the Harvey–Bradshaw Index (HBI) and the Simple Endoscopic Score for Crohn’s Disease (SES-CD) were reduced significantly from baseline (p = 0.01 for both), whereas the median value of the Lémann score did not change significantly (4.9 vs 4.2, p = 0.28). Based on univariate analysis, only penetrating behaviour at baseline was more likely to be associated with the risk of surgery (OR 5.09; CI 95% 1.45 to 17.82; p = 0.01). Survival curve analysis and Cox regression model showed that patients treated with infliximab were more likely to avoid surgery compared with those treated with adalimumab (HR 3.13, CI 95% 1.63–12.67, p = 0.0038 and OR 6.54; CI 95% 0.66–3.09; p = 0.0026, respectively). The use of infliximab and longer duration of anti-TNF therapy were protective from surgery (OR 1.05; CI 95% 0.01–0.08; p = 0.0017). Conclusion Half of CD patients starting anti-TNF therapy for stricturing disease avoided surgery after a median follow-up of 3 years. Penetrating behaviour was associated with an increased risk of surgery. Patients treated with infliximab remained free of surgery longer than patients treated with adalimumab did.