Matters close to the heart: adaptation of a tacrolimus model to inform tacrolimus therapy using the PRIOR approach
Ranita Kirubakaran (1,2), David W. Uster (3), Stefanie Hennig (4,5), Jane E. Carland (1,2,6), Richard O. Day (1,2), Sebastian G. Wicha (3), Sophie L. Stocker (1,2,7).
(1) St Vincent's Clinical School, Faculty of Medicine and Health, The University of New South Wales, Sydney, Australia. (2) Department of Clinical Pharmacology & Toxicology, St Vincent's Hospital Sydney, Australia. (3) Department of Clinical Pharmacy, Institute of Pharmacy, University of Hamburg, Germany. (4) Certara, Inc., Princeton, New Jersey, USA. (5) School of Clinical Sciences, Faculty of Health, Queensland University of Technology, Brisbane, Australia. (6) School of Medical Sciences, Faculty of Medicine and Health, The University of New South Wales, Sydney, Australia. (7) School of Pharmacy, Faculty of Medicine and Health, The University of Sydney, Australia.
Introduction. Tacrolimus is the cornerstone immunosuppressant used to prevent and treat graft rejection in heart transplant (HTX) recipients. Due to the considerable pharmacokinetic variability and narrow therapeutic range, monitoring of tacrolimus trough concentrations is recommended to individualise tacrolimus therapy. However, the collection of inappropriately timed concentrations (i.e., not trough concentration) hampers the successful implementation of trough concentration-guided tacrolimus dosing in routine clinical practice. This is evident from a recent exploration of tacrolimus monitoring practices at St Vincent’s Hospital Sydney which highlighted that only 25% of tacrolimus concentrations collected were steady-state trough concentrations [1]. Given that a Bayesian dosing software may assist in the interpretation of tacrolimus concentrations collected at any time within a dosing interval, the predictive performance of a commercially available Bayesian dosing software approved for use in Australia was assessed [2]. Unfortunately, the software could provide acceptable predictions of tacrolimus concentrations only after 11 days of tacrolimus therapy. The delay in the achievement of acceptable predictions was due to the reported instability of tacrolimus pharmacokinetics in the early post-transplant phase and stabilisation of the clinically significant tacrolimus-azole antifungal interaction.
Growing literature suggest that population pharmacokinetic model-guided dosing can individualize tacrolimus therapy in transplant recipients [3] and holds great promise to improve patient outcomes [4]. Hence, we externally evaluated numerous published tacrolimus population pharmacokinetic models (n=17) [5] selected from our recent systematic review [6] for clinical implementation. Unfortunately, none of the models adequately described tacrolimus disposition in HTX recipients, both with and without concomitant azole antifungal therapy [5].
Objective. To develop and evaluate a population pharmacokinetic model for tacrolimus in HTX recipients receiving and not receiving concomitant azole antifungal therapy.
Methods. Retrospective data of patients who underwent heart transplantation in 2017 and 2018 at St Vincent’s Hospital Sydney and received the oral immediate-release formulation of tacrolimus (Prograf®) were obtained up to 391 days post-HTX. Due to data sparseness, values from a tacrolimus population pharmacokinetic model published by Sikma et al. [7] were used to stabilise the estimation of pharmacokinetic parameters (Ka, V1, Q, V2) during model development. The published model was selected because it has previously displayed the best bias and imprecision in predicting tacrolimus concentrations in HTX recipients receiving concomitant azole antifungal therapy [5]. The published model, which did not include covariates was implemented in NONMEM v7.4.3. Subsequently, the model was tested on data from HTX recipients in 2018 (1138 concentrations) using the PRIOR NWPRI subroutine. The population average oral bioavailability (F) was fixed to one (1) [7]. All between-occasion variabilities (BOV) on CL, V1, Ka and F reported in the published model were fixed to zero during the estimation step to reduce model flexibility. The published model was extended to quantify the effect of concomitant azole antifungal therapy on tacrolimus apparent clearance (CL/F). Internal and external evaluations of the updated model were assessed using prediction-corrected visual predictive check (pcVPC, PsN v5.0.0). The pcVPC graphics were stratified based on azole antifungal use. External evaluation was performed using data from HTX recipients in 2017 (1369 tacrolimus concentrations).
Results. The pharmacokinetic parameter estimates of the updated model were close to the values of the published model except for Ka (0.579 vs. 0.295 h-1; published vs. updated model) and CL/F (19.6 vs. 16.8 L/h). Concomitant azole antifungal reduced tacrolimus CL/F by 79%. Between-subject variability (BSV) on CL/F for HTX recipients receiving concomitant azole antifungal therapy was 92% and while not receiving concomitant azole antifungal therapy was 68%. Internal and external evaluation showed that the updated model adequately described the data, for both with and without concomitant azole antifungal therapy.
Conclusions. The updated model was able to describe tacrolimus pharmacokinetics in HTX recipients, with and without concomitant azole antifungal therapy. Prospective evaluation is planned to access the clinical utility of the updated model.
References:
[1] Kirubakaran et al (2018) Tacrolimus dosing and monitoring in heart transplant: a retrospective observational study [Poster Presentation]. ASCEPT Scientific Meeting, Adelaide, Australia.
[2] Kirubakaran et al (2019). Predictive performance of a Bayesian forecasting software for tacrolimus in adult heart transplant [Oral Presentation]. PAGANZ, Auckland, New Zealand.
[3] Brunet et al (2019) Ther Drug Monit 41(3), 261−307.
[4] Storset et al (2015) Transplantation 99(10), 2158−2166.
[5] Kirubakaran et al (2019) External evaluation of population pharmacokinetic models of tacrolimus in adult heart transplant recipients [Oral Presentation]. ASCEPT-PAGANZ Joint Scientific Meeting, Queenstown, New Zealand.
[6] Kirubakaran et al (2020) Clin Pharmacokinet 59(11), 1357−1392.
[7] Sikma et al (2020) Eur J Drug Metab Pharmacokinet 45(1), 123−134.