PKP-009 Evaluation of a population pharmacokinetic model of infliximab in rheumatoid arthritis for prediction of individual dosage requirements

Background Infliximab (IFX) is a chimeric antitumour necrosis factor α monoclonal antibody used in rheumatoid arthritis (RA). It shows large interindividual variability in serum concentrations during treatment. The use of previously developed population pharmacokinetic (PPK) models might help to guide dosing recommendations to improve response. External validation provides the most compelling evidence for the validity of a PPK model. Purpose To evaluate a previously developed PPK model for IFX in RA patients using an external population dataset. Material and methods RA patients receiving IFX between July 2014 and July 2015 were included. Pre-dose serum concentrations (C min ) (mg/L) and antibodies against IFX (ATI) were determined at steady state by enzyme linked immunosorbent assay (Promonitor). Demographic, biochemical and haematologic covariates, disease activity score (DAS28), tender joints (tender28) and swollen joints (swollen28) were recorded. The PPK model reported by Ternant et al . 1 was implemented in NONMEM v.7.2 and used as a Bayesian predictor. Bias and imprecision were estimated by the median prediction error and the root median squared prediction error, respectively. Prediction errors were calculated for each individual predicted concentration. Results 36 C min values from 17 patients were available (14 women; 6 concomitant psoriasis). IFX was administered at 3 mg/kg/8 weeks (n = 13), 3 mg/kg/10 weeks (n = 2), 3 mg/kg/6 weeks (n = 1) or 5 mg/kg/8 weeks (n = 2). Concurrent immunomodulatory therapy was given to 15 patients. The median C min (range) was 0.29 mg/L (0.01–3.87 mg/L). Median (range) values for the most relevant characteristics were: C reactive protein 3.5 mg/dL (0.5–60.7); serum albumin 43 g/L (38–48); leucocytes 7.15 × 10 9 /L (3.6–12.3); erythrocyte sedimentation rate 10 mm (2–76); DAS28 2.26 (0.63–4.71); swollen28 0 (0–8); and tender28 0 (0–8). Two patients developed ATI. Bias and imprecision estimated values were -0.03 (95% CI −0.53 to 0.43 mg/L) and 0.10 (95% CI 0.01 to 0.67 mg/L, 24%), respectively. Conclusion Acceptable bias and imprecision values were provided by the PPK model of Ternant et al . 1 but a further evaluation using larger datasets will be required to confirm these results. References and/or Acknowledgements Ternant D, Ducourau E, Perdriger A, et al . Relationship between inflammation and infliximab pharmacokinetics in rheumatoid arthritis. BJCP 2013;78:118-28 No conflict of interest.