Introduction

The feasibility of anti-TNF discontinuation in inflammatory bowel disease (IBD) must be proven in clinical trials including patients in clinical, endoscopic, and radiologic remission at the time of anti-TNF withdrawal, to give recommendations for clinical practice.

Aims & Methods

Aims: Primary: to compare the rates of clinical remission at 1 year in patients who discontinue anti-TNF treatment vs. those who maintain it. Secondary objectives: to know the effect of anti-TNF withdrawal on relapse-free time, mucosal healing and safety; and to identify predictive factors for relapse.
Methods: Prospective, quadruple-blind, multicentre, randomised, controlled trial. Patients with ulcerative colitis (UC) or Crohn’s disease (CD) in clinical remission for > 6 months were randomised to maintain anti-TNF treatment [maintenance arm (MA)] or to withdraw it [withdrawal arm (WA)]. Patients who were on infliximab (IFX) received IFX 5 mg/kg or an intravenous placebo every 8 weeks, while patients on adalimumab (ADA) received subcutaneous ADA 40 mg or placebo every other week. Patients were followed-up until month 12 or up to the time of clinical relapse, whichever came first. Inclusion and exclusion criteria were as follows: patients should have been in clinical remission for ≥6 months at a non-intensified dose of anti-TNF (where an intensified dose is defined as ≥10 mg/kg/8 weeks or 5 mg/kg/≤4 weeks for IFX, and 40 mg/week for ADA); patients should have been receiving concomitant immunosuppressants; for patients with CD, the indication for anti-TNF treatment must have been for luminal involvement only (not perianal); the baseline colonoscopy performed within 3 months of inclusion should have not shown significant lesions; for patients with ileal or ileocolic CD, the magnetic resonance enterography performed within 3 months of inclusion should have not shown significant lesions; patients with prior intestinal resection were excluded. Results were analysed by intention-to-treat (ITT) and by per-protocol (PP). Local investigators were blinded to faecal calprotectin (FC) and IFX and ADA trough levels. On-site monitoring was performed to assess data quality.

Results

159 patients were screened, from whom 140 were randomised and comprised the ITT cohort: 70 allocated to the MA and 70 to the WA. Fifteen patients dropped out before the end of follow-up (12 months or relapse), leaving 63 patients in the MA and 62 patients in the WA for the PP analysis. The characteristics of patients in the MA and WA were similar. The proportion of patients who maintained clinical remission [59/70 (84%), 95% confidence interval (CI)=74-92% in the MA vs. 53/70 (76%), 95%CI=64-85% in the WA] and those with significant endoscopic lesions at the end of follow-up (8% in the MA vs. 19% in the WA, p>0.05) were similar between groups. Only the proportion of patients with FC >250 mg/g was higher in the WA at the end of follow-up (12% in the MA vs. 33% in the WA, p<0.01). The proportion of patients with sustained clinical remission during follow-up was no different between groups. In the multivariate analysis, FC >250 mg/g at baseline, but not anti-TNF withdrawal, was the only variable associated with lower likelihood of sustained clinical remission (HR=0.2, CI95%=0.07 - 0.7). The same percentage of patients in both groups had at least one adverse event (69%). The proportion of patients with serious adverse events was also similar between groups (4% in MA vs. 7% in WA).

Conclusion

Anti-TNF withdrawal in selected IBD patients in clinical, endoscopic, and radiologic remission could be feasible without an increase in the risk of clinical relapse.