Model-based dosing of Anti-Thymocyte Globulin in children

Background

ATG is shown to impact GvHD and GF, with varying impact on outcome and survival (Finke Lancet Onc 09, Walker Lancet Haem 20, Kroger NEJM 16, Soiffer JCO 17). This may be due to ATG-induced poor T-cell reconstitution after HSCT, potentially leading to decreased graft-versus-leukaemia effect and viral reactivations. Although ATG is widely used in allogeneic HCT, its use in this setting is not mentioned in the summary of product characteristics.

ATG Thymoglobulin versus ATG Grafalon

The two most commonly used types of ATG in HCT are Thymoglobulin and Grafalon. These are not biosimilar, and therefore differ in biologically active fraction (capacity to bind to human targets). In recent years, model-based individualized dosing has been developed for Thymoglobulin (see below). For Grafalon, no published population pharmacokinetic models and/or evidence for a therapeutic window is available (results have recently been presented at meetings). As such, while Thymoglobulin can be dosed using the model-based dosing algorithm (see above), no such dosing is available for Grafalon. Hopefully, model-based dosing for Grafalon will be available in future.

Highly variable pharmacokinetics: fixed dosing leads to highly variable exposure

A population pharmacokinetic model (Admiraal Clin PK 2015) describes the kinetics of ATG (Thymoglobulin) in children, and can make accurate predictions of exposure based on body weight and lymphocyte counts. The model shows that a cumulative dose of 10mg/kg leads to large differences in exposure (Figure 1).

Figure 1: Model-based simulation results of active Thymoglobulin concentrations showing the population predictions of three representative individuals receiving Thymoglobulin according to the current dosing regimen (cumulative dose of 10 mg/kg in 4 consecutive days), weighing 5 kg (a), 20 kg (b), and 40 kg (c). Adapted from Admiraal et al, Clin PK 2015

Exposure to ATG is highly related to clinical outcomes

In retrospective analyses in multiple cohorts (Admiraal Lancet Haem 2016, Admiraal Blood, Lakkaraja Blood Adv 21), ATG exposure is highly related to outcome. High exposure to ATG before graft infusion prevents GvHD and graft failure, while exposure to ATG after graft infusion (ie exposure of the graft to ATG) should be low to promote T-cell recovery (>50/uL CD4 within 100 days after transplant; CD4IR). Optimal exposure of ATG after graft infusion predicts OS; the optimum differs per cell source: for cord blood the exposure should be very low (<20 AU*day/mL), while for BM/PBSC this should be <50 AU*day/mL (Figure 2).

Figure 2: Overall survival according to ATG exposure after graft infusion in cord blood with a cut-off at 20 AU*day/mL (panel A) and bone marrow with a cut-off at 50 AU*day/mL (panel B). Adapted from Admiraal et al, Lancet Haem 2016

Early T-cell recovery is a strong predictor for survival

Multiple analyses (Admiraal Lancet Haem 2016, Admiraal Blood, de Koning Blood 21, v Roessel Cytother 21, Troullioud Lucas Cytother 2023) in different cohorts underlined the importance of early T-cell recovery in improving therapy-related mortality.

Model-based dosing leads to improved T-cell recovery, OS and TRM

Results from 2, 3 and 4 were combined to develop a model-based (MB) dosing (MBD) algorithm for ATG aiming for optimal exposures that would lead to potent T-cell recovery and reduced mortality. The PARACHUTE-trial (Admiraal, Lancet Haem 22) evaluated the MB-dosing in a single-arm open label phase II trial, powered for T-cell recovery. The trial showed a marked improvement in successful T-cell recovery from 50% (historical controls) to 75% of pts. No increase in GvHD was observed.

We combined the recent analysis of the 5-year follow-up results of the PARACHUTE combining with real-world data using the same regimen (patients from Utrecht and New York; manuscript in preparation, presented at ASTCT Tandem 2024). Here, MB-dosing improves OS as compared to historical controls (Figure 3). The incidence of cGvHD and GF is markedly lower with MB-dosing. Importantly, successful CD4IR was strongly associated with improved TRM.

Figure 3: Updated results of PARACHUTE with real-world data (Model-based dosing) versus historical controls (Fixed dosing). Panel A: overall survival according to treatment group; panel B: transplant related mortality according to treatment group; panel C: transplant related mortality according to CD4 T-cell recovery (CD4IR) and treatment group. Confidential data, submitted

Trial experience with model-based dosing of ATG

Currently, the dosing nomogram has been used in 1) the PARACHUTE-trial (see above), 2) PRAISE-IR (NY, USA, currently recruiting) investigating individualized dosing in ex-vivo CD34 selected grafts and 3) SCRIPT-AML (multinational RCT, recruiting) investigating BuCyMel vs BuFluClo with a center-based choice to use either ATG Grafalon (fixed dosing) or ATG Thymoglobulin (individualized dosing).

Dosing tables

Please note that separate tables are used for either Bone Marrow/Peripheral Blood or for Cord Blood, in which the dose is selected using body weight, lymphocyte counts and stem cell source. The dosing tables are designed for pediatric patients receiving bone marrow, peripheral blood or cord blood transplant (not including haplo-identical settings), receiving myeloablative conditioning.

How to implement ATG model-based dosing in your centre

Model-based dosing is easy to implement

Support in implementing model-based dosing

Medical Responsibility

Background

Supporting papers