Prediction of neutrophil nadir and recovery following paediatric haematopoietic cell transplantation with busulfan conditioning.

Abstract

Aims: In haematopoietic cell transplantation (HCT), neutropenia resulting from myelosuppression is an expected endpoint following busulfan-based conditioning. However, if prolonged, neutropenia can lead to complications like serious infection and death. The routine use of pharmacokinetic (PK)-guided busulfan dosing has decreased toxicities; however, patients still have serious, sometimes fatal, complications of drug-induced neutropenia. We sought to investigate whether the time-course of neutropenia after HCT could be predicted for paediatric patients following busulfan-based conditioning. Such predictions could guide care, such as timing of infectious prophylaxis and/or growth factor administration.

Methods: This was a single-centre, retrospective study of 146 patients with malignant or nonmalignant disorders treated with allogeneic or autologous HCT. An advanced PKPD model of neutrophil dynamics post-HCT was built that included two parallel neutrophil maturation pathways for host and donor cells, expanded transit compartments for neutrophil maturation, cell number-based feedback loops to the proliferating compartment, GCSF effects, and direct/indirect busulfan killing effects.

Results: The model predicted neutrophil recovery well using patient data beyond Day +21 post-HCT. When using patient data prior to Day +21 post-HCT, neutrophil recovery was less accurate (as measured by the sum of the absolute neutrophil count prediction error) due to unpredictable complications influencing neutrophil counts. Four clinical cases illustrate the strengths and challenges of the model.

Conclusions: This study suggests real-time incorporation of patient-specific data through Day +21 post-HCT is required to predict neutrophil dynamics following neutrophil nadir in paediatric HCT.