Detection of myelodysplastic syndrome/ acute myeloid leukemia evolving from aplastic anemia in children, treated with recombinant human G-CSF

Backgrounds and Objectives: Recombinant human granulocyte colony-stimulating factor (G-CSF) has clear benefits in patients with severe neutropenia. However, recent reports of myelodysplastic syndrome/acute myeloid leukemia (t-MDS/AML) developing after treatment with immunosuppressants and G-CSF has raised concern over the use of this agent in patients with aplastic anemia. Design and Methods: We undertook a multi-institutional, non-randomized study of 112 children given a diagnosis of aplastic anemia, and then treated with different immunosuppressants with or without G-CSF. In each case, bone marrow specimens were tested at study entry and every 6 months for 3 years to detect t-MDS/AML, defined by stringent morphological and molecular/cytogenetic criteria. Incidence rates were calculated by the person-years statistical method. Results: As of December 2001, all eligible patients had been followed for a median of 3 years, and the G-CSF (+) group had received a median total G-CSF dose of 30,100 micrograms altogether, administered over a median of 4 months. Only one case of MDS developed among the G-CSF (+) patients (n=81), compared with three in the group receiving other agents (n=31). This isolated case was not associated with monosomy 7, the cytogenetic abnormality most often linked to G-CSF treatment. Incidence rates of MDS in the two groups were not significantly different (3.8 vs. 22.4 per 1,000 patient-years at risk, p=0.125). There were no cases of overt AML in either cohort. Interpretation and Conclusions: G-CSF therapy did not increase the risk of t-MDS/AML development in children with aplastic anemia over a median follow-up of 3.7 years