Sine Buhl [1], Maria Rasmussen [1], Jørn Brynskov [1], Mark A Ainsworth [1], Klaus Bendtzen [2], Pia H Klausen [1], Nils Bolstad [3], David J Warren[3], Casper Steenholdt [1].

Affiliates

[1] Dept. of Gastroenterology, Copenhagen University Hospital Herlev, Denmark, [2] Institute for Inflammation Research, Rigshospitalet, Copenhagen, Denmark, [3] Department of Medical Biochemistry, Oslo University Hospital, Oslo, Norway

Abstract

Background: Primary non-response to infliximab (IFX) induction therapy inherits a poor prognosis with high risk of surgery in patients with inflammatory bowel disease (IBD). We explored underlying mechanisms and therapeutic thresholds in an effort to provide basis for optimizing therapy by therapeutic drug monitoring (TDM).

Methods: Prospective cohort study including 166 IBD patients (CD n=93; UC n=73) having received standard IFX induction (5 mg/kg at weeks 0-2-6). Response was assessed at week 14 and supported by disease activity indices in 50%. Trough IFX and anti-IFX antibodies (Abs) were assessed at weeks 2 (n=148), 6 (n=108), and 14 (n=66). In a partly overlapping cohort, circulating TNFα was measured in matched primary non-responders (n=29) and responders (n=21).

Results: 18 patients (11%) had primary IFX failure. IFX was consistently lower throughout the induction phase in primary non-responders (Week 2: IFX median 18.9 μg/mL vs. 23.3, p<0.05. Week 6: 8.4 vs. 17.0, p<0.05. Week 14: 0.5 vs. 5.6, p<0.001). Therapeutic IFX thresholds during induction phase were 22.9 μg/mL at week 2 (sensitivity 51%, specificity 80%, AUCROC 0.67, p<0.05), 11.8 at week 6 (72%, 77%, 0.71, p<0.05;), and 1.4 at week 14 (91%, 75%, 0.85, p<0.01). Anti-IFX Ab-formation occurred in 28% of primary non-responders (n=5) and associated with low IFX and treatment failure (OR 13.7 [2.8-67.5], p<0.01), supporting an immune-mediated pharmacokinetic (PK) failure mechanism. Markers of disease activity (disease activity scores, albumin, CRP) also associated with low IFX supporting non-immune-mediated PK failure at high inflammatory load. In ulcerative colitis (n=28), TNFα was higher throughout induction in primary non-responders (Week 6: TNFα median 13.5 pg/mL vs. 9.2, p<0.01. Week 14: 18.5 vs. 9.6, p<0.05), indicating insufficient inhibition of TNFα mediated activity. Conversely, similar TNFα levels in non-responders and responders with Crohn’s disease (n=22) (Week 6: 11.5 vs. 12.5, p0.05. Week 14: 10.9 vs. 10.2, p0.05) indicated effective TNFα inhibition in non-responders and supported predominantly pharmacodynamic (PD) failure due to non-TNFα-driven disease pathways.

Conclusion: PK- and PD problems are common causes of primary IFX failure. Prospective testing of TDM-based induction strategies utilizing therapeutic thresholds to address underlying mechanistic issues is warranted.