Sicilian Network for Inflammatory Bowel Disease (SN-IBD). A propensity score-matched comparison of infliximab and adalimumab in naïve and non-naïve patients with Crohn’s disease.

Background: In the absence of head-to-head trials, there is an unmeet need to better understand the relative effectiveness of different biologics in inflammatory bowel disease (IBD). The Sicilian Network for Inflammatory Bowel Disease (SN-IBD) is a group composed by all Sicilian centres which continuously enter in a web-based software all clinical data of IBD patients treated with biologics. Methods: Data of all incident Crohn’s disease (CD) patients treated with infliximab (IFX) and adalimumab (ADA) from January 2013 to April 2017 were extracted from the cohort of SN-IBD. Patients were divided in biologic-naïve and non-naïve, and the two groups were analysed singularly. We used a one-to-two propensity score matching (1 IFX: 2 ADA) accounting for the main baseline characteristics in naïve patients, and a one-to-one propensity score matching (1 IFX: 1 ADA) in non-naïve. Results: Seven hundred and forty-seven naïve and 188 non-naïve patients were included. After propensity score matching, 453 naïve (IFX: 151; ADA: 302) and 100 non-naïve patients (total treatments: 122; IFX: 61; ADA: 61) were analysed. Among naïve patients, the rates of response/remission at 12 weeks for IFX and ADA were 80.1% and 81.1%,, respectively (adjusted OR 0.97, p = 0.923); over a median follow-up of 11.8 months, the rates of response/remission for IFX and ADA were 70.2% and 66.2%, respectively, without significant differences (adjusted OR 1.14, p = 0.401). Among non-naïve patients, the rates of response/remission at 12 weeks for IFX and ADA were 68.9% and 60.7%, respectively (adjusted OR 1.54, p = 0.320); over a median follow-up of 8.9 months, the rates of response/remission for IFX and ADA were 57.4% and 54.1%, respectively, without significant differences (adjusted OR 1.96, p = 0.297). Cox regression analysis showed no differences in risk of treatment failure between ADA and IFX, neither in naïve (adjusted HR 1.23, p = 0.381) nor in non-naïve patients (adjusted HR 1.23, p = 0.488). At multivariable conditional logistic regression analysis of naïve CD patients, upper GI involvement (OR 0.18, p = 0.038), previous surgery (OR 0.24, p = 0.003), and older age (OR 0.97, p = 0.036) were associated with lower clinical benefit at 12 weeks, while previous surgery was the only independent predictor of treatment failure at the end of follow-up (HR 2.13, p = 0.03). Mixed effect Cox analysis showed that non-naïve patients experiencing more than one previous line of treatment with biologics have a significant higher risk of treatment failure compared with those previously treated with one biologic only (HR 2.57, p = 0.002) Conclusions: In this large, propensity score matched, real-life, multicentre, cohort study of CD patients, there was no significant difference in the effectiveness of ADA and IFX. Both drugs showed a good efficacy.