Background and Objective Mycophenolic acid (MPA) exhibits considerable inter-individual variability in drug exposure, which can result in acute graft rejection as well as hematological or infectious adverse effects. Evidence from our group and others indicates that certain uremic toxins may contribute to this variability through pharmacokinetic (PK) interactions. This study aimed to develop a novel population PK (popPK) model to investigate how conjugated metabolites of p-cresol and indole (i.e., toxicokinetically important uremic toxins) affect total MPA PK, and to conduct model-based simulations to identify potentially relevant dosing recommendations. Methods A prospective observational study enrolled adult kidney transplant recipients on steady-state oral mycophenolate mofetil (MMF; prodrug of MPA) with tacrolimus (±prednisone). Total plasma concentrations of p-cresol sulfate (pCS), p-cresol glucuronide (pCG), indoxyl sulfate (IxS), indoxyl glucuronide (IxG), MPA, and its major glucuronide metabolite (MPAG) were quantified with our validated liquid chromatography tandem-mass spectrometry assays. PopPK modelling was conducted with stochastic approximation expectation-maximization, and Monte-Carlo simulation was used to assess the potential impacts of significant covariates on MPA exposure. Results Forty-one participants contributed 283 samples across three early post-transplant periods (~1, ~3, and ~6 months). The final popPK model was described by first-order absorption (Ka = 0.672 [0.47–0.99] h−1, estimate [95% confidence interval]) with lag time (Tlag = 0.403 [0.39–0.42] h), two compartments for MPA (central volume, Vc = 1.09 [0.75–1.53] L; peripheral volume, Vp = 113.9 [76.33–197.33] L; intercompartmental clearance, Q = 15.9 [10.11–25.61] L/h; and clearance = fixed at 1.4 L/h), and a single compartment for MPAG (clearance, CLMPAG = 0.296 [0.23–0.35] L/h; MPA-to-MPAG metabolic conversion, Kpm = 3.21 [2.46–4.18] h−1). A proportional error model with inter-individual and inter-occasional variability best described the random effects. Potentially significant covariates were “pCS exposure” on MPA Tlag, Kpm, and Q (covariate coefficients, β = − 0.226 [−0.53 to 0.079], −0.133 [−0.25 to −0.033], and −0.162 [−0.39 to 0.13], respectively); “IxS exposure” and “estimated glomerular filtration rate (eGFR)” on CLMPAG (β = −0.181 [−0.28 to −0.035] and 0.407 [0.085–0.73], respectively); and “IxG exposure” on MPA Tlag (β = 0.295 [−0.0057 to 0.59]). The model was validated by goodness-of-fit plots, residual plots, visual-predictive checks, and non-parametric bootstrapping. Model simulations identified pCS as a covariate positively influencing total MPA exposure; that pCS and eGFR had negative effects on MPAG exposure, potentially opposing the effects of IxS; whereas IxG had no effect on either MPA or MPAG. Conclusion To our knowledge, this is the first popPK model to mechanistically characterize PK interactions between uremic toxins and total MPA in kidney transplant recipients. Our findings indicate that each toxin has distinct interaction effects, with pCS emerging as potentially relevant. Additional investigations are required to elucidate the clinical impacts of the identified toxin-MPA PK interactions in this population.

Impacts of Uremic Toxins on the Population Pharmacokinetics of Total Mycophenolic Acid and its Glucuronide Metabolite in Adult Kidney Transplant Recipients

期刊论文
发表日期：

2026年5月1日

查看全文

Ala’a R. Al-Dajani,
Yan Rong,
Jinal Adhiya,
Puja Dhungana,
Patrick Mayo,
Penny Colbourne,
Sita Gourishankar,
Tony K. L. Kiang

摘要

Background and Objective

Mycophenolic acid (MPA) exhibits considerable inter-individual variability in drug exposure, which can result in acute graft rejection as well as hematological or infectious adverse effects. Evidence from our group and others indicates that certain uremic toxins may contribute to this variability through pharmacokinetic (PK) interactions. This study aimed to develop a novel population PK (popPK) model to investigate how conjugated metabolites of p-cresol and indole (i.e., toxicokinetically important uremic toxins) affect total MPA PK, and to conduct model-based simulations to identify potentially relevant dosing recommendations.

Methods

A prospective observational study enrolled adult kidney transplant recipients on steady-state oral mycophenolate mofetil (MMF; prodrug of MPA) with tacrolimus (±prednisone). Total plasma concentrations of p-cresol sulfate (pCS), p-cresol glucuronide (pCG), indoxyl sulfate (IxS), indoxyl glucuronide (IxG), MPA, and its major glucuronide metabolite (MPAG) were quantified with our validated liquid chromatography tandem-mass spectrometry assays. PopPK modelling was conducted with stochastic approximation expectation-maximization, and Monte-Carlo simulation was used to assess the potential impacts of significant covariates on MPA exposure.

Results

Forty-one participants contributed 283 samples across three early post-transplant periods (~1, ~3, and ~6 months). The final popPK model was described by first-order absorption (K_a = 0.672 [0.47–0.99] h⁻¹, estimate [95% confidence interval]) with lag time (T_lag = 0.403 [0.39–0.42] h), two compartments for MPA (central volume, V_c = 1.09 [0.75–1.53] L; peripheral volume, V_p = 113.9 [76.33–197.33] L; intercompartmental clearance, Q = 15.9 [10.11–25.61] L/h; and clearance = fixed at 1.4 L/h), and a single compartment for MPAG (clearance, CL_MPAG = 0.296 [0.23–0.35] L/h; MPA-to-MPAG metabolic conversion, K_pm = 3.21 [2.46–4.18] h⁻¹). A proportional error model with inter-individual and inter-occasional variability best described the random effects. Potentially significant covariates were “pCS exposure” on MPA T_lag, K_pm, and Q (covariate coefficients, β = − 0.226 [−0.53 to 0.079], −0.133 [−0.25 to −0.033], and −0.162 [−0.39 to 0.13], respectively); “IxS exposure” and “estimated glomerular filtration rate (eGFR)” on CL_MPAG (β = −0.181 [−0.28 to −0.035] and 0.407 [0.085–0.73], respectively); and “IxG exposure” on MPA T_lag (β = 0.295 [−0.0057 to 0.59]). The model was validated by goodness-of-fit plots, residual plots, visual-predictive checks, and non-parametric bootstrapping. Model simulations identified pCS as a covariate positively influencing total MPA exposure; that pCS and eGFR had negative effects on MPAG exposure, potentially opposing the effects of IxS; whereas IxG had no effect on either MPA or MPAG.

Conclusion

To our knowledge, this is the first popPK model to mechanistically characterize PK interactions between uremic toxins and total MPA in kidney transplant recipients. Our findings indicate that each toxin has distinct interaction effects, with pCS emerging as potentially relevant. Additional investigations are required to elucidate the clinical impacts of the identified toxin-MPA PK interactions in this population.