BACKGROUND: Anti-thymocyte globulin (ATG) prevents graft-versus-host disease (GVHD) in allogeneic hematopoietic cell transplantation (HCT). However, variable ATG exposure impacts transplant outcomes.
OBJECTIVE: We aimed to develop a population pharmacokinetic (popPK) model for ATG in myeloablative HCT and evaluate the relationship between area under the time-concentration curve (AUC) and mortality to identify the optimal ATG AUC.
METHODS: We studied 200 adult HCT recipients who received myeloablative conditioning (MAC) and a peripheral blood stem cell graft from 7/8 or 8/8 HLA-matched related or unrelated donors. ATG was given on days -2, -1, and 0. All patients received additional GVHD prophylaxis with methotrexate and cyclosporine. Serum concentration of lymphocyte-binding ATG was determined by flow cytometry in 2,140 samples. For the popPK modeling, the cohort was split into a model development cohort (n = 134) and a validation cohort (n = 66). The modeling was performed using Monolix Suite 2024R1. The relationship between model estimated AUCs and mortality was evaluated in all 200 patients (combined development and validation subcohorts) using a Cox proportional hazards model. The relationship between model estimated AUCs and cause-specific outcomes (e.g., relapse or acute graft-versus-host disease [aGVHD]) was evaluated using a competing risk analysis.
RESULTS: A two-compartment model with parallel linear and target-mediated elimination best described ATG disposition. Population means and residual standard errors (RSE%) were 11.78 L (2.05%) for the central volume of distribution (V1), 0.20 L/h (4.36%), and 2.20 U/L (7.08%) for the initial ATG-binding capacity of lymphocytes in the central compartment. Lean body weight (LBW) positively correlated with V1, CL, and intercompartmental clearance, while pre-ATG absolute lymphocyte count (ALC) positively correlated with R0_initial. Internal and external validation (using the development and validation subcohorts, respectively) confirmed model stability and robustness. The optimal ATG AUC range was 30-45 U·day/L. In multivariate analysis, patients whose AUC was within this range had lower mortality than those whose AUC was outside this range (hazard ratio=0.46, p = 0.03). The low mortality of patients with AUC within the range of 30-45 U·day/L appeared to be due to both low incidence of grade III-IV aGVHD, which was high in patients with AUC <30 U·day/L, and low incidence of relapse, which was high in patients with AUC >45 U·day/L.
CONCLUSION: This novel model described the pharmacokinetics of ATG in adult HCT recipients following MAC. The model identified LBW and ALC as significant covariates for ATG disposition. Furthermore, we identified the optimal ATG AUC (associated with the lowest mortality). These findings provide the foundation for developing an individualized dosing strategy aimed at improving post-HCT survival.