Ruben Faelens (1,4), Erwin Dreesen (2), Gert Van Assche(3), Marc Ferrante(3), Séverine Vermeire(3), Ann Gils (2), Thomas Bouillon (1)
(1) Drug Delivery and Disposition, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven – University of Leuven, Leuven, Belgium (2) Therapeutic and Diagnostic Antibodies, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven – University of Leuven, Leuven, Belgium (3) Department of Gastroenterology and Hepatology, University Hospitals Leuven, Leuven, Belgium (4) SGS Exprimo, Mechelen, Belgium
Introduction Infliximab (IFX) is used in patients with ulcerative colitis (UC) to achieve mucosal healing (MH, defined as Mayo endoscopic sub-score <= 1). A pharmacokinetic/pharmacodynamic (PKPD) model was established [1] linking IFX exposure to probability of mucosal healing (pMH), based on a previously published dataset [2]. The model allows to optimize dosing to increase population exposures, which is predicted to improve clinical outcomes. Compared to open loop dosing, therapeutic drug monitoring (TDM) may be used to optimize dosing even further.
Objective To evaluate different dosing regimens towards obtaining mucosal healing, and the potential benefit of TDM-based dose adjustment.
Methods For simulation, the base model without covariates on PK was used: a 1cpt model with between-subject variability (BSV) and within-subject variability (WSV) on Ke (CVs%=39; 21.5) and V (CVs%=42; 14.5). The PD model predicts pMH based on individual cumulative area under the curve (CAUC) at day 84 and baseline score. For targeted dosing, a pMH of 70%, predicted to correspond to a CAUC of 3400µg/mLday or 4800µg/mLday for a baseline score of 2 or 3 respectively was aimed for.
The following dosing regimen were evaluated (dosing at weeks 0, 2 and 6):
- 3x5mg/kg as per the label.
- 3x10mg/kg.
- Fixed dose based on baseline score (open loop). 3x600mg and 3x800mg is predicted to achieve pMH=70% for a baseline of 2 and 3 respectively.
- As (3), with further adjustment at day 14 using Bayesian estimation of BSV based on serum concentration (closed loop), and selecting the dose that is predicted to reach individual pMH of 70%.
- As (4), assuming perfect knowledge of BSV.
To evaluate probability of study success (PoSS), a chi-squared test on the number of patients achieving MH was used to compare both arms. Simulations were implemented using Simulo Expert v7.2. TDM was implemented using the TDMore framework, available at https://github.com/rfaelens/tdmore .
Results Reported values are mean (for dose or pMH) or median (for CAUC) values [90% prediction interval]
| Scenario | Average Dose (mg) | CAUC (µg/mL*day) for baseline score of 2 (target=3400) | CAUC (µg/mL*day) for baseline score of 3 (target=4800) | pMH (%) | |
| (1) | 5mg/kg | 360 | 2162 [811,5521] | same | 54 [27,77] |
| (2) | 10mg/kg | 720 | 4314 [1626,11077] | same | 70 [46,87] |
| (3) | Open loop | 701 | 3676 [1457,8920] | 4878 [1932,11899] | 70 [48,86] |
| (4) | Closed loop | 697 | 3527 [2519,5235] | 4941 [3534,7277] | 71 [63,78] |
| (5) | Perfect knowledge of BSV | 701 | 3425 [2638,4483] | 4837 [3718,6305] | 70 [64,76] |
| Scenario | Arm 1 | Arm 2 | Patients per arm | PoSS |
| Dose-controlled trial | 5mg/kg | 10mg/kg | 100 | 57% |
| 167 | 80% | |||
| Exposure-controlled trial | pMH = 54% | pMH = 70% | 100 | 60% |
| 156 | 80% |
Discussion The pMH with the current regimen is 54% and confirmed in our simulations. Increasing the dose to 3x10mg/kg is predicted to achieve pMH of 70%. Dichotomisation of fixed dose by baseline Mayo score does not reduce the distribution of individual pMH, due to high BSV.
Closed loop dosing is predicted to reduce BSV of pMH. Note that this beneficial effect is not detectable on a population level, since the fraction cured in the population remains unchanged compared to a fixed dose regimen yielding identical average exposure. Any dose saved in over-exposed patients is equally consumed in under-exposed patients, resulting in identical average consumption of drug on a population level.
As N=167 patients per arm are needed to achieve a PoSS (power) of 80% at alpha=0.05, studies evaluating a 10mg/kg dose vs a 5 mg/kg dose should be adequately powered. An exposure-controlled trial does not greatly improve PoSS when evaluating efficacy of 5mg/kg vs 10mg/kg.
Conclusion Assuming PK drives PD, higher induction doses than 5mg/kg should be explored, as they are predicted to have better outcomes. Compared to 5mg/kg, 10mg/kg is predicted to raise average exposure from 2150 µg/mLday to 4300 µg/mLday, and pMH from 54% to 70%. TDM aiming for a target exposure increases pMH for under-exposed patients, but decreases the pMH for over-exposed patients. Population pMH is therefore predicted to not improve using this technique, although individual probabilities of MH are equalized. Finally, an exposure-controlled trial is not expected to improve the power of studies using the MH dichotomous endpoint.
References:
[1] Dreesen et al. “Development of a population pharmacokinetic and pharmacodynamic model to describe the effect of infliximab induction therapy on mucosal healing in patients with ulcerative colitis” submitted, PAGE 2018
[2] Papamichael, Konstantinos, et al. “Infliximab concentration thresholds during induction therapy are associated with short-term mucosal healing in patients with ulcerative colitis.” Clinical Gastroenterology and Hepatology 14.4 (2016): 543-549.
[3] Abrantes, Jönsson, Karlsson, Nielsen “Handling inter-occasion variability in model-based therapeutic drug monitoring”, PAGE 2017
Reference: PAGE 27 (2018) Abstr 8529 [www.page-meeting.org/?abstract=8529]
Poster: Drug/Disease Modelling - Other Topics