Immunomodulator withdrawal from anti-TNF therapy is not associated with loss of response in inflammatory bowel disease

Abstract

BACKGROUND AND AIMS: The benefit of concomitant immunomodulators (thiopurines or methotrexate) in patients with inflammatory bowel disease (IBD) on anti-tumor necrosis factor α (anti-TNF) (infliximab or adalimumab) maintenance therapy is debated. We compared outcomes after immunomodulator withdrawal vs continuation of combination therapy. METHODS: This was a retrospective cohort study in a general hospital and a tertiary referral center. We included adult IBD patients, receiving anti-TNF therapy for ≥4 months, plus an immunomodulator at baseline, between January 1, 2011, and January 1, 2019. The primary endpoints were loss of response (LOR) (ie, anti-TNF discontinuation because of disease activity) and anti-drug antibodies. Adjusted hazard ratios (aHRs) were calculated by mixed-effects Cox regression analysis. RESULTS: We included 614 treatment episodes of combination therapy in 543 individuals, yielding 1664 patient-years of follow-up. The immunomodulator was withdrawn in 296 (48.2%) episodes after 0.9 (interquartile range, 0.6-2.1) years, which was not associated with a higher risk of LOR (aHR, 1.08; 95% confidence interval [CI], 0.72-1.61), although anti-drug antibodies were detected more frequently (aHR, 2.14; 95% CI, 1.17-3.94), compared with continuation. Clinical remission at the time of withdrawal reduced the risk of LOR (aHR, 0.48; 95% CI, 0.25-0.93), while longer duration of combination therapy before withdrawal decreased the risk of anti-drug antibodies (HR per year, 0.56; 95% CI, 0.32-0.91). Higher prewithdrawal infliximab trough levels reduced the subsequent risks of anti-drug antibodies and LOR. Infliximab trough levels were lower after immunomodulator withdrawal (P = .01). CONCLUSIONS: Patients who withdrew the immunomodulator in this retrospective cohort were not at increased risk of LOR within the following 1-2 years, but an increase in anti-drug antibodies was observed. Our findings require prospective validation, preferably in adequately powered randomized controlled trials.