BACKGROUND Discontinuation of anti-TNF therapy in patients with inflammatory bowel diseases (IBD) in remission remains a controversial issue. The aims of our study were to assess the proportion of patients who relapse after cessation of biological treatment, and to identify potential risk factors of disease relapse. METHODS Consecutive IBD patients who discontinued anti-TNF therapy in steroid-free clinical and endoscopic remission were prospectively followed. Multiple logistic regression and Cox proportional-hazards models were used to assess the predictors of disease relapse. RESULTS Seventy-eight IBD patients (Crohn's disease, CD 61; ulcerative colitis, UC 17) were included and followed for a median of 30 months (range 7-47). A total of 32 (53%) CD patients and nine (53%) UC patients relapsed by the end of the follow-up with a median time to relapse of 8 months (range 1-25) in CD patients and 14 months (range 4-37) in UC patients, respectively. The cumulative probabilities of maintaining remission at 6, 12, and 24 months were 82%, 59%, and 51% in CD patients, and 77%, 77%, and 64% in UC patients, respectively. Survival of CD patients who were in deep remission (clinical and endoscopic healing; faecal calprotectin <150 mg/kg; CRP ≤5 mg/l) was not better compared with those who did not fulfill these criteria. In multivariate models, only colonic CD protected patients from disease relapse. CONCLUSIONS Approximately half of the IBD patients relapsed within 2 years after anti-TNF discontinuation. In CD patients, no difference between those who were or were not in deep remission was found. Colonic localization protected patients from relapse.