Clinical and cytogenetic responses to granulocyte-macrophage colony-stimulating factor in therapy-related myelodysplasia

We treated 10 patients with a therapy-related myelodysplastic syndrome with escalating doses of granulocyte-macrophage colony-stimulating factor (GM-CSF; sargramostim) in a phase II trial and used sequential cytogenetic analyses to determine whether there was stimulation of nonclonal hematopoiesis. The GM-CSF was administered by continuous intravenous infusion over 2 hours daily for 14 days, followed by a 14-day rest period. The initial starting dose was 60 micrograms/m2/d. The GM-CSF dose was escalated within individual patients to 125 micrograms/m2, 250 micrograms/m2, and then 500 micrograms/m2/d until the peripheral blood neutrophil count at least doubled and exceeded 1,000/microL. GM-CSF treatment then continued in monthly maintenance cycles. During 57 treatment courses, the neutrophil count increased in 52 but only doubled and exceeded 1,000/microL in 21. Mild eosinophilia was stimulated in five patients, but only two had greater than 1,000 eosinophils/microL. In only three patients was any stimulation of platelet or red blood cell production observed, and thus, little change in transfusion requirements occurred. The bone marrow karyotypes from individual patients either remained completely abnormal or became increasingly abnormal over the course of treatment. We found no evidence that GM-CSF preferentially stimulated normal marrow stem cells to proliferate or had the ability to eradicate the cytogenetically abnormal clone by inducing terminal differentiation. Although the effect on granulopoiesis was transient and dependent on continued GM-CSF treatment, the increase in the neutrophil count was clinically important in some patients, allowing more effective control of ongoing infections.