A repeated time-to-event model linking infliximab clearance to the risk of relapse during maintenance treatment of patients with inflammatory bowel disease
Zhigang Wang 1, Niels Vande Casteele 2, Marc Ferrante 3,4, Séverine Vermeire 3,4, Erwin Dreesen 1
1 Department of Pharmaceutical and Pharmacological Sciences, KU Leuven, Leuven, Belgium; 2 Department of Medicine, University of California San Diego, San Diego, USA; 3 Department of Gastroenterology and Hepatology, University Hospitals Leuven, Leuven, Belgium; 4 Department of Chronic Diseases and Metabolism, KU Leuven, Leuven, Belgium
Objectives:
Therapeutic drug monitoring (TDM) of intravenously administered infliximab is common practice in patients with inflammatory bowel disease (IBD). Dosing is optimised towards a trough concentration (TC) target of 5 mg/L, which is generally believed to result in an optimal therapeutic response.[1] However, infliximab concentrations do not only drive response, but they are also driven by disease activity (e.g. increased infliximab clearance due to leakage through the inflamed bowel wall). This bi-directional relation between pharmacokinetics (PK) and pharmacodynamics (PD) may undermine the clinical impact of TDM as adequate exposure in the presence of high infliximab clearance may still result in poor outcomes. This study aimed to disentangle the bi-directional PK-PD relationship of infliximab and investigate the potential benefit of clearance monitoring over concentration monitoring.
Methods:
Data from 189 patients (65 ulcerative colitis [UC], 124 Crohn’s disease [CD]) who underwent infliximab dose optimisation based on clinical features (n=78) or TDM (n=101) were included.[2] All patients were in steroid-free combined clinical and biological remission at the start of the study and remission was systematically evaluated at each hospital visit over one year. Clinical remission was defined as a Partial Mayo Score (PMS) <3 and no subscore >1 in patients with UC, and Harvey-Bradshaw Index (HBI) <5 in patients with CD. Biological remission was defined as C-reactive protein (CRP) <10 mg/L.
Time-to-relapse was modelled using a parametric repeated time-to-event (RTTE) model. Various hazard functions (exponential, Weibull, Gompertz, Surge) were explored. Time-varying covariates, including infliximab TC, the presence of anti-infliximab antibodies, infliximab clearance, the infliximab dose amount, the infliximab dosing interval, average infliximab daily dose per interval, serum albumin, CRP, haemoglobin, white blood cell count, and body weight, as well as baseline covariates such as disease type and sex, were investigated as predictors of the relapse hazard risk. Infliximab clearance was calculated using weighted averaging of the Bayesian forecasted estimates of four population PK models.[3] NONMEM 7.5 (Icon Plc, Gaithersburg, MA) with the Laplace approximation was used to obtain the likelihood.
Results:
Of all patients, 49.7% (n=27 [UC]; n=67 [CD]) experienced at least one relapse event during the year following start of dose optimisation, with up to four and nine relapse events for patients with UC and CD, respectively. An exponential hazard model best described the relapse data of patients with UC, with a constant baseline hazard of 0.00519 day-1 (RSE=9.7%). The Gompertz hazard model best described the data of patients with CD, with the baseline relapse hazard increasing over time from 0.00519 day-1 at inclusion to 0.0162 day-1 at one year (hazard ratio [HR]=3.12; 95% CI [2.30-4.24]). The infliximab TC did not have a significant impact on the relapse hazard risk (P>0.05). However, the infliximab clearance had a significant impact on the baseline hazard (dOFV 16.4 points). An increase in estimated infliximab clearance from 0.371 (25th percentile) to 0.561 (75th percentile) L/day was associated with an HR of 1.38 (95% CI [1.15-1.66]). An increase in measured CRP from 1.40 (25th percentile) to 12.8 (75th percentile) mg/L was associated with an HR of 2.24 (95% CI [1.84-2.71]). The functions of the two final hazard models are presented below:
UC: 0.00519 * exp[1.7*(clearance-0.453) + 0.0706*(CRP-4.7)]
CD: 0.00519 * exp[1.7*(clearance-0.453) + 0.0706*(CRP-4.7)] * exp(0.00312*time)
Conclusions:
We developed an RTTE model that demonstrated a significant association between higher infliximab clearance and increased risk of relapse in patients with IBD. While infliximab TCs lose their predictive ability for remission when doses are optimised, the infliximab clearance remains a relevant predictor of remission. This also explains why patients with adequate exposure still lose response. Our infliximab clearance monitoring software tool can provide valuable information to forecast the risk of relapse when patients with IBD are on TDM.
References:
[1] Vande Casteele et al. Gastroenterology (2017): 153, 835-857. [2] Vande Casteele et al. Gastroenterology (2015): 148, 1320-1329. [3] Kantasiripitak et al. CPT:PSP (2022): 11, 1045–1059.