Abstract Background Combination therapy with anti-TNF compounds and immunomodulators (IMM; thiopurine or methotrexate) is superior to IMM or anti-TNF monotherapy in patients with inflammatory bowel disease (IBD). IMMs are frequently discontinued during the maintenance phase to mitigate the risk of adverse events, but long-term consequences of this practice are not well studied. We explored the real-world outcomes after IMM discontinuation, including loss-of-response (LOR), dose escalations, immunogenicity and trough levels. Methods This was a multicenter, retrospective cohort study in a general hospital and a tertiary referral center. We included adult patients with IBD, treated >4 months with infliximab (IFX) or adalimumab (ADA) and an IMM at baseline between 2011–2019. The IMM had to be started within 30 days of anti-TNF initiation, or continued for >30 days in case of prior IMM monotherapy. LOR was defined as anti-TNF discontinuation due to disease activity. Adjusted hazards rates (aHR) were calculated using mixed-effects Cox regression analysis with time-varying covariates, accounting for follow-up prior to and after IMM cessation. We adjusted for sex, age, BMI, smoking, Crohn’s disease (CD) vs ulcerative colitis (UC), disease duration, primary sclerosing cholangitis, rheumatological comorbidity, ADA vs IFX, and prior anti-TNF exposure. Results We included 615 episodes of combination therapy (543 individual patients; CD, n=382, 70%). The IMM was discontinued in 296 (48%) episodes after a median of 0.9 (IQR: 0.6 – 2.1) years, at which point 252 (85%) patients were in clinical remission. IMM withdrawal was performed as part of a de-escalation strategy (n=158, 53%), for intolerance (n=86, 29%) or for miscellaneous reasons (n=52, 18%). During a median follow-up of 1.7 (IQR 0.8 – 3.5) years after IMM withdrawal, 46 (16%) patients experienced LOR, 79 (32%) required dose-escalation and 31 (10.3%) developed anti-drug antibodies. Compared to IMM continuation, withdrawal did not significantly increase the risk of LOR (aHR 1.10, 95%CI: 0.74 – 1.64), but more patients required dose escalations (aHR 1.42, 95%CI 1.02 – 1.97) or developed anti-drug antibodies (aHR 2.22, 95%CI 1.21 – 4.08). Among patients who stopped the IMM, clinical remission at IMM withdrawal was the only predictor of LOR (aHR 0.48, 95%CI 0.23 – 0.99), while higher BMI (aHR 1.09, 95%CI 1.01 – 1.17) and shorter duration of combination therapy (Figure 1, aHR 0.57 per year, 95%CI 0.33 – 0.96) increased the risk of immunogenicity. IFX, but not ADA, trough levels decreased significantly after IMM withdrawal. Conclusion Withdrawal of immunomodulators is not associated with higher risk of LOR, but does increase the risk of dose-escalation and unfavorable pharmacokinetics.
Read full abstract