(S-103) Model-Informed Oral Tacrolimus Dose Optimization for Adult Recipients of Allogeneic Hematopoietic Cell Transplantation Receiving Post-Transplant Cyclophosphamide

Riley Randolph – University of North Carolina at Chapel Hill; Tyler Dunlap, PharmD, DABCP – University of North Carolina at Chapel Hill, Metrum Research Group; Stephany Gonzalez Tineo – University of North Carolina at Chapel Hill; Ryan Kemper – University of North Carolina at Chapel Hill; Jing Zhu, PharmD, PhD – University of North Carolina at Chapel Hill; Yu Fei Wang – University of North Carolina at Chapel Hill; Eric Weimer, PhD – University of North Carolina at Chapel Hill; Anson Snow, MD – UNC Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center; Paul Armistead, MD, PhD – University of North Carolina at Chapel Hill, UNC Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center; Daniel Weiner, PhD – University of North Carolina at Chapel Hill; J. Ryan Shaw, PharmD, BCOP – University of North Carolina at Chapel Hill, University of North Carolina Medical Center; Daniel Crona, PharmD, PhD – University of North Carolina at Chapel Hill, UNC Lineberger Comprehensive Cancer Center, University of North Carolina Medical Center

Objectives: Tacrolimus is the cornerstone of acute graft-versus-host disease (aGVHD) prophylaxis among allogeneic hematopoietic cell transplants (allo-HCT), but PK variability and a narrow therapeutic index complicate dosing. Previously, we developed a population pharmacokinetic (popPK) model with clinically distinct subpopulations based on chemotherapy conditioning regimen and CYP3A5 metabolizer phenotype. The model was used to propose a dosing strategy to better achieve tacrolimus trough concentrations (Cmin,ss) within the University of North Carolina Medical Center (UNCMC) institutional target range (ITR; 5-10 ng/mL). The model excluded patients receiving post-transplant cyclophosphamide (PTCy), which could represent its own clinically distinct subpopulation. This study evaluated the applicability of our previous popPK model in the PTCy population and attainability of ITR under the current dosing regimen.

Methods: Adults who received their first allo-HCT at UNCMC, PTCy, and oral tacrolimus for aGVHD prophylaxis were eligible (UNC IRB 22-2292). Tacrolimus Cmin,ss, obtained from standard-of-care therapeutic drug monitoring, were collected. A simulation-based approach was used to assess to what extent observed tacrolimus Cmin,ss in PTCy subjects differed from concentrations simulated by the previous model. The average ratio of simulated to observed Cmin,ss was summarized across 1,000 Monte Carlo replicates and the FDA clinical equivalence standard (0.8 to 1.25) was used to determine adequacy of the previous model in PTCy subjects. Then, observed data from the prior model and PTCy data were pooled to refit the model, evaluating with goodness-of-fit and simulation-based diagnostics. Individual empirical Bayesian estimates of PK parameters from PTCy subjects were used to simulate tacrolimus Cmin,ss with the current UNCMC weight-based dosing strategy. Proportions of subjects with predicted Cmin,ss within the ITR and below the ITR ( < 5 ng/mL) were evaluated.

Results: 556 tacrolimus Cmin,ss from 89 adult allo-HCT recipients receiving PTCy were analyzed. Most subjects were diagnosed with acute myeloid leukemia (55%) and received reduced intensity conditioning (RIC) chemotherapy (72%). A mix of CYP3A5 phenotypes were observed: poor metabolizers (PM; 55%) and intermediate/normal metabolizers (IM/NM; 45%). The median [90% prediction interval (PI)] of the average simulated to observed Cmin,ss ratio was 0.64 [0.58, 0.70], aligning PTCy subjects with those in the previous model treated with RIC. The model was updated to include a new RIC/PTCy covariate effect on clearance. Subsequent simulations estimated a median [90% PI] of the average simulated to observed Cmin,ss ratio of 1.08 [0.99, 1.18] and was validated by visual predictive checks. Of the predicted Cmin,ss values in PTCy subjects, 32% were within the UNCMC ITR but 30% were still subtherapeutic.

Conclusions: The previous model underpredicted tacrolimus Cmin,ss in PTCy patients, but the pooled popPK analysis adequately described the tacrolimus PK in allo-HCT patients and suggests a RIC-like effect of PTCy on oral tacrolimus clearance. Tacrolimus dose adjustments may be needed in IM/NM patients receiving PTCy to prevent subtherapeutic Cmin,ss.

Citations: [1] Bolaños-Meade, J., Hamadani, M., Wu, J., Al Malki, M. M., Martens, M. J., Runaas, L., Elmariah, H., Rezvani, A. R., Gooptu, M., Larkin, K. T., Shaffer, B. C., El Jurdi, N., Loren, A. W., Solh, M., Hall, A. C., Alousi, A. M., Jamy, O. H., Perales, M. A., Yao, J. M., Applegate, K., … BMT CTN 1703 Investigators (2023). Post-Transplantation Cyclophosphamide-Based Graft-versus-Host Disease Prophylaxis. The New England journal of medicine, 388(25), 2338–2348. https://doi.org/10.1056/NEJMoa2215943
[2] Dunlap, et al. The first de novo oral tacrolimus population pharmacokinetic model for adult allogeneic hematopoietic cell transplant recipients provides clinical opportunities for precision dosing. Provisionally accepted to Clinical Pharmacokinetics in February 2025. Under revision.

Keywords: Tacrolimus, post-transplant cyclophosphamide, pharmacokinetics