High efficacy of fludarabine-containing therapy (FLAG-FLANG) in poor risk acute myeloid leukemia

Background. Elderly patients with acute myeloid leukemia (AML), those refractory to induction chemotherapy and those with so-called secondary leukemia have unfavorable prognoses and require innovative therapeutic approaches. Fludarabine allows an increased accumulation of Ara-CTP in leukemic cells and inhibits DNA repair mechanisms; therefore its association with Ara-C and mitoxantrone results in a synergistic effect. Materials and Methods. From May 1993 to February 1996, fludarabine-containing regimens (FLAG and FLANG) were employed as induction therapy in 51 high-risk AML patients. Diagnosis of AML in 22 patients was preceded by a myelodysplastic syndrome lasting more than six months; 8 of the 29 de novo AML cases (28%) were refractory to previous chemotherapy, 9 (31%) were treated for early relapse, 12 (41%) presented poor prognostic factors at diagnosis. The median age was 64 (range 33-76) years and the FAB subtypes were the following: M0 3, M1 5, M2 28, M4 7, M5 8. Forty-eight per cent of patients showed poor prognosis chromosomal abnormalities. FLAG (24 patients) consisted of both fludarabine 30 mg/sqm over 30 minutes followed 4 hours later by Ara-C 2 g/sqm over 4 hours (for 5 days) and G-CSF 300 mu g/day administered 12 hours before fludarabine, for a total of 5 doses. FLANG (27 patients) had a shorter duration (3 days), reduced Ara-C dosage (1 g/sqm) and administration of mitoxantrone (10 mg/sqm) at the end of Ara-C infusion. Results. Recovery of both neutrophils (PMN > 0.5x10(9)/L) and platelets (Fit > 20x10(9)/L) required a median of 16 days from the end of therapy. Overall, 30 patients (59%) achieved CR, 6 (11%) PR and 10 (20%) were refractory; 5 (10%) experienced early death (cerebral hemorrhage or infection). The length of complete response ranged from 2 to 26 months with a median follow-up of 8 months. De novo and secondary AML registered 62 and 54% CR rates, respectively. Eight out of 10 patients refractory to conventional schemes achieved CR (80%) but only 3 out of 10 treated for relapse obtained CR (30%). Conclusions. FLAG and FLANG showed similar activity and toxicity while proving to be highly effective and relatively well-tolerated treatments for high-risk de novo AML. Secondary leukemias seemed to be responsive as well, but the presence of an unfavorable karyotype alteration lowered the response rate.