P589 A population pharmacokinetic model to improve mucosal healing upon golimumab induction therapy in patients with ulcerative colitis
W. Kantasiripitak1, E. Dreesen*1, I. Detrez1, S. Stefanović2, D. Drobne2, S. Vermeire3,4, M. Ferrante3,4, A. Gils1
1University of Leuven, Department of Pharmaceutical and Pharmacological Sciences, Leuven, Belgium, 2University Medical Centre Ljubljana, Department of Gastroenterology and Hepatology, Ljubljana, Slovenia, 3University Hospitals Leuven, Department of Gastroenterology and Hepatology, Leuven, Belgium, 4University of Leuven, Department of Chronic Diseases, Metabolism and Ageing, Leuven, Belgium
From the PURSUIT programme, it is known that golimumab (GLM) trough concentrations (TC) >2.5 mg/l at week (w)6 of induction therapy are associated with clinical response in patients with ulcerative colitis (UC).1 No TC threshold has been established for mucosal healing (MH; Mayo endoscopic score ≤1). A population pharmacokinetic (popPK) model may support dose optimisation to improve attainment of a predefined TC target.
GLM concentration–time data of 56 patients with UC (335 venepuncture and 296 dried bloodspot (DBS) samples) were obtained from 2 study centres (University Hospitals Leuven, Belgium and Ljubljana University Medical Centre, Slovenia).2,3 A popPK model was developed using NONMEM 7.4. Exposure during GLM induction therapy was linked to MH at w14.
A two-compartment popPK model with linear absorption and elimination showed good predictive capacity (Figure 1A). The estimated popPK parameters (typical value [%RSE]) were absorption constant (0.511 day−1 [8%]), apparent clearance CL/F (0.407 l/day [6%]), volume of distribution in the central compartment (9.16 l [5%]) and peripheral compartment Vp/F (3.21 l [22%]) and inter-compartmental clearance (0.464 l/day [13%]). Antibodies to GLM and higher alkaline phosphatase increased GLM CL/F, while prior biological use was associated with a larger Vp/F, all predicting lower GLM exposure. Still, 48% and 147% of the interindividual variability (IIV) on CL/F and Vp/F remained unexplained. A total of 14/40 patients (35%, 16/56 no endoscopy data available) achieved MH after GLM induction therapy. These patients had higher model-predicted GLM TC at w6 (median 7.6 mg/l, interquartile range [5.8–8.0]) compared with patients not achieving MH (4.7 mg/l [3.3–6.8]) (
With the currently approved induction dosing of GLM, only 10/40 (25%) reached the proposed 7.4 mg/l TC target at w6, suggesting underexposure in a substantial proportion of patients. This popPK model shows good predictive capacity and may be implemented in a therapeutic drug monitoring software tool to allow better targeting of the here established exposure targets (TC and AUC) in individual patients. Still, Bayesian updating of individuals’ PK parameters using early DBS samples is recommended given the remaining large IIV.
(A) Prediction-corrected visual predictive check. (B) Density plot and receiver-operating characteristic (ROC) curve of pGLM TC at w6 as a predictor of MH at w14.
1. Adedokun OJ, Xu Z, Marano CW,
2. Detrez I, Dreesen E, Van Assche G,
3. Stefanovic S, Detrez I, Compernolle G,