Zeinab Daher Abdi (1), Annick Rousseau (1,2), Aurélie Prémaud (1,2)
(1) INSERM, UMR-850, limoges, France, (2)Univ Limoges, Limoges, France.
Objectives: This study aimed to investigate, in adult kidney transplant recipients, the relationship between longitudinal exposure to co-administered immunosuppressive drugs (i.e mycophenolic acid (MPA) and an calcineurin inhibitor (either cyclosporine or tacrolimus) and a composite efficacy endpoint including acute rejection, graft loss and death.
Methods: Data from 222 patients were analyzed: 23 events were observed in 126 patients receiving cyclosporine against 15 in 96 patients receiving tacrolimus (p=0.61) in the two first years post-transplantation. Within NONMEM V7.2.0, the longitudinal drug exposures described using mixed-effects models were incorporated as time-dependent covariates in a parametric time-to-event model. An interval-censored approach was used for patients who had experienced an event to take into account that rejection occurred in a time interval prior to the time of diagnosis. Donor and recipient characteristics were tested as covariates.
Results: A sigmoid Emax model was selected to describe the time course of MPA exposure and exponential models were retained for the time course of cyclosporine and tacrolimus. Model-predicted time-varying distributions of MPA, cyclosporine and tacrolimus were in good agreement with observed values. The developed time-to-event model showed that the studied efficacy composite outcome was associated to longitudinal exposure to MPA and to onset of cytomegalovirus infection or disease. Within the observed ranges, calcineurin inhibitor exposures were not significantly associated with efficacy (i.e. acute rejection, graft loss and death). The risk of acute rejection, graft loss or death adjusted on the longitudinal calcineurin inhibitor co-exposures, decreased by 4% (95% confidence interval Hazard Ratio 0.93-0.99) per 1 mg.h/L increase of MPA area under the curve. The onset of cytomegalovirus infection/disease significantly increased this risk (Hazard Ratio=10.9; 95% confidence interval 6.5-21.7). Visual-predictive-check and Kaplan-Meier plots showed that the developed time-to-event model described well the rejection-free and failure-free graft survival curves.
Conclusions: For the first time, a time-to-event model that considers the combined longitudinal exposures to two co-administered drugs was developed. This work advocates for the avoidance of unnecessary high calcineurin inhibitors dosing and puts forward new arguments for MPA concentration monitoring.
Reference: PAGE 24 () Abstr 3373 [www.page-meeting.org/?abstract=3373]
Poster: Methodology - New Modelling Approaches