Abstract • 25

Background: Over the last decade recombinant human granulocyte-colony stimulating factor (G-CSF) has had a major impact on management of "severe chronic neutropenia", a term used collectively that refers to congenital, idiopathic or cyclic neutropenia. More tan 90% of patients respond to G-CSF with an increase in neutrophils to >1.5×109/l, and with a marked reduction in infections and a vastly improved survival. Several reports indicate that responders with congenital have developed myelodysplastic syndromes and acute myeloblastic leukemia (MDS/AML) raising the question of the role of G-CSF in the pathogenesis of the malignant transformation.

Methods: The Severe Chronic Neutropenia International Registry (SCNIR) has collected data of 531 neutropenic patients from 13 countries treated with G-CSF from 1987 to 1997, including 302 patients with congenital neutropenia. To clarify the risk of MDS/AML in congenital neutropenia and the potential role of G-CSF in pathogenesis, these data were analyzed with respect to treatment and patient demographics.

Results: The 302 patients with congenital neutropenia were observed on G-CSF treatment for a mean of five years (range 0.1 to 10 years). Twenty-seven developed MDS/AML giving a crude rate of malignant transformation of nine percent. Twenty-six had Kostmann's type of congenital neutropenia and one had Shwachman-Diamond syndrome. None of 229 patients with idiopathic or cyclic neutropenia for whom similar data were maintained developed MDS/AML. Conversion to MDS/AML was frequently associated with acquired marrow cytogenetic clonal changes: 14 developed partial or complete loss of chromosome 7, and nine manifested abnormalities of chromosome 21 (usually trisomy 21). The annual rate for MDS/AML for each yearly interval was less than two percent per year of treatment. No statistically significant relationship between age at onset of MDS/AML and patient gender, dose of G-CSF or duration of treatment with G-CSF was found (p values for all >0.15).

Conclusion: In the pre-cytokine era there were reports of MDS/AML developing spontaneously in patients with congenital forms of neutropenia indicating a propensity for malignant myeloid transformation. Our data do not support a cause-and-effect relationship between development of MDS/AML and G-CSF or duration of therapy in congenital neutropenia, nor with any other patient demographics. It is likely that the improved survival of congenital neutropenia patients with G-CSF therapy allows time for the expression of the leukemic predisposition that characterizes the natural history of these disorders.