Granulocyte-Colony-Stimulating Factor (G-CSF) Accelerates Normal Myeloid Recovery More Efficiently in Patients with Favorable Than with Intermediate or Unfavorable Cytogenetic Risk Acute Myeloid Leukemias (AML): A Retrospective Analysis of the East German Hematology Oncology Study Group (OSHO).

Objectives: Hematopoietic growth factors are essential in accelerating myeloid reconstitution and reducing treatment morbidity of hematological and non-hematological malignant diseases. In order to understand in more detail the effect of G-CSF on myeloid regeneration, patients (pts) with AML (excluding APL) enrolled in the OSHO AML96 and AML97 trials were analyzed with respect to cytogenetic risk factor, G-CSF treatment and myeloid recovery.

Methods: A total of 603 eligible pts from 19 centers were included in this retrospective study. The diagnosis of AML was centrally reviewed and cytogenetic analyses performed in accredited laboratories. Pts were grouped in favorable (ie CBF leukemias: t(8;21); inv16) intermediary (normal or others) and unfavorable (−5/5q-; −7/7q-; abn11q23; inv3, or complex) cytogenetic risks. Treatment consisted of 8gm/m² Ara-C for induction (1gm/m² q12h or 2gm/m²/d on d1,3,5,7) in combination with an anthracycline (idarubicine 12mg/m²/d in patients <60a or mitoxantrone 10mg/m²/d on d1-3 in patients >60a). G-CSF was started on day 10 after chemotherapy and continued until myeloid reconstitution in 394 pts from 10 centers.

Results: After induction chemotherapy, G-CSF accelerated leukocyte recovery (>1x10⁹/L) by 3 days (median 21d vs 24d; n=560). The effect of G-CSF was, however, more evident in the favorable cytogenetic risks (19d vs 25d; p<0.0001), followed by the intermediate (21d vs 24d; p<0.0001) and not significant in the unfavorable (22d vs 25d; p=0.29). In pts without G-CSF treatment, myeloid reconstitution was identical in the 3 cytogenetic groups (25d vs 24d vs 25d). Subsequently, only pts with complete remission after the first induction treatment were selected (n=302) and the duration of leukopenia analyzed with respect to G-CSF and cytogenetic findings. Again, G-CSF shortened the duration of leukopenia in all 3 cytogenetic groups, and the effect was clearly influenced by cytogenetics. The median duration to myeloid recovery in pts with CBF AML was 25d without G-CSF and 19d with G-CSF treatment (p=0.0001). Pts with intermediate cytogenetics recovered after a median of 24d without and after 21d with G-CSF (p= 0.0003). The outcome in pts with unfavorable karyotype was similar: 25d without and 21d with G-CSF treatment, respectively (p=0.004). However, whereas no difference in myeloid reconstitution was observed between the 3 cytogenetic groups without G-CSF administration, clear differences were seen in G-CSF treated pts between the CBF AML and the others (6d vs 3-4d; p<0.0007). In a multivariate analysis, the cytogenetic risk group (p=0.002), remission state after first induction (p=0.004), and G-CSF administration (p<0.0001) independently proved to influence myeloid regeneration.

Conclusion: We conclude that the accelerating effect of G-CSF on myeloid recovery correlates with cytogenetic findings and is greater in CBF AML than in intermediary or unfavorable cytogenetics. The mechanisms underlying this effect are currently unknown, but might involve increased susceptibility of normal hematopoiesis to G-CSF in CBF AML, or might rely on the differentiating effect of G-CSF on CBF AML.