P502 Application of an algorithm-based precision-dosing model to a real-world cohort of patients on infliximab maintenance therapy: implications for drug usage and cost

Nguyen, A.(1,2,3)*;Gibson, P.(1,2);Upton , R.(4,5);Mould, D.(4);Sparrow, M.P.(1,2);

(1)Monash University, Central Clinical School, Melbourne, Australia;(2)Alfred Health, Department of Gastroenterology, Melbourne, Australia;(3)Charité – Universitätsmedizin Berlin, Medical Department of Gastroenterology- Infectiology and Rheumatology, Berlin, Germany;(4)Projections Research Inc., Projections Research Inc., Phoenixville, United States;(5)University of South Australia, Clinical and Health Sciences, Adelaide, Australia;


Infliximab is registered for use in patients with inflammatory bowel disease (IBD) at an intravenous maintenance dose of 5 mg/kg eight-weekly. Secondary loss of response during maintenance therapy occurs in up to 50% of patients. New precision dosing models, which forecast infliximab doses to achieve targeted trough levels, reduce non-response and improve patient outcomes. Aim: To compare standard 5 mg/kg eight-weekly infliximab dosing to the infliximab doses theoretically required to achieve predefined infliximab trough target levels, as determined by the iDOSE® precision dosing dashboard model.


IBD patients treated with eight-weekly maintenance infliximab therapy dosed at 5 mg/kg were included. A recorded infliximab dose was considered dose X if the following was available: three previous infliximab doses, albumin and C-reactive protein at time of dose X, trough infliximab level immediately preceding dose X and the patient’s weight.  The Baysient LLC iDOSE® software was then used to predict dose X to maintain therapeutic infliximab trough levels of ≥ either 5-10 or 3-7 µg/ml until the next eight-weekly dose. The predicted dose X was compared to the actual dose X given.


174 patients (56% male) were included with 417 dose X recordings. Of those, 135 had Crohn’s disease, 31 ulcerative colitis and 8 IBD-unclassified. Median age was 36 (IQR 29-47) years with a disease duration of 5 (1-13) years. 76% of patients were receiving immunomodulators and 6% were smokers. Duration of infliximab therapy before dose X was 2 (0-4) years. Trough infliximab levels were 5.0 (2.5-7.6) µg/ml, serum albumin was 38 (35-40) g/L and C-reactive protein 3 (3-4) mg/L. Comparing actual dose X with predicted dose X, 52% and 32% of doses were subtherapeutic when aiming for trough levels of 5-10 and 3-7 µg/ml respectively. The underdosing occurred by 5.6 and 5.3 mg/kg per dose, respectively. Cost analysis showed total costs increased by 102% and 29% when aiming for trough levels 5-10 and 3-7 µg/ml. On multivariable regression analysis, significant associations with subtherapeutic infliximab levels were ulcerative colitis compared to Crohn’s disease [OR: 9.81, 95% CI: 1.28-75.40, p = 0.028] and pre-dose X infliximab trough level [OR: 0.07, 95% CI: 0.03-0.15, p < 0.001].


More than half of maintenance infliximab drug doses given at our institution were too low to achieve therapeutic infliximab blood levels of at least 5 µg/ml. These patients would likely benefit from an increased dose or reduced dosing interval of infliximab. These findings need to be confirmed prospectively.