OBJECTIVES: I. Assess the efficacy of recombinant human granulocyte colony-stimulating factor (G-CSF) in raising the absolute neutrophil count, platelet count, and hemoglobin level in patients with inherited bone marrow failure syndromes.

II. Assess the efficacy of a reduced maintenance dose in patients who respond to daily G-CSF.

III. Assess the toxic effects of G-CSF in these patients. IV. Measure bone marrow progenitor colonies before and after G-CSF. V. Measure CD34-positive cells in marrow and blood before and after G-CSF using flow cytometry and immunohistochemistry.

PROTOCOL OUTLINE: Patients receive granulocyte colony-stimulating factor (G-CSF) subcutaneously every day for 8 weeks; nonresponders receive an increased dose for an additional 8 weeks. Patients who respond at week 8 or 16 are then tapered to a lower maintenance dose of G-CSF administered every other day through week 40. The dose is adjusted to maintain an absolute neutrophil count above 1500.

Patients are removed from study for failure to achieve a complete response by week 16, unacceptable nonhematologic toxicity, the identification of a clonal karyotype in marrow, or the onset of leukemia.

PROTOCOL ENTRY CRITERIA:

--Disease Characteristics--

Inherited bone marrow failure syndrome, including:

• Fanconi's anemia
• Dyskeratosis congenita
• Shwachman syndrome
• Amegakaryocytic thrombocytopenia
• Decreased megakaryocytes in infancy
• No thrombocytopenia with absent radius syndrome (TAR)
• No trisomy 13 or 18
• No clonal bone marrow karyotype

--Prior/Concurrent Therapy--

• At least 4 weeks since growth factors
• Concurrent therapy allowed if not altered for 30 days prior to entry through week 8
• No concurrent investigational drugs

--Patient Characteristics--

• Hematopoietic: ANC <1000
• No leukemia
• Other: No medical or psychiatric contraindication to protocol participation
• No pregnant or nursing women