Phase I Trial of Clofarabine and High Dose Cytarabine with G-CSF Priming (G-CLAC) for Relapsed or Refractory Acute Myeloid Leukemia

Background: Clofarabine is active in both relapsed and newly-diagnosed AML, and recent data suggest that addition of low dose cytarabine (ara-C) improves efficacy for induction of patients age 60 or older (Faderl et al Blood epub June 18, 2008). Clofarabine has also been combined with ara-C at a dose of 1 g/m² daily for five days, and this regimen resulted in a 52% complete remission rate for patients age 50 or older undergoing induction for AML (

Objective: We conducted a phase 1 trial to determine the maximum tolerated dose (MTD) of clofarabine, and the dose-limiting toxicities (DLTs) of the combination of clofarabine and high dose ara-C with G-CSF priming, in the treatment of patients with relapsed or refractory AML.

Methods: Patients (pts) ages 18–70 with relapsed or refractory AML were included. Eligibility included up to two attempts of prior induction for the current relapse or new diagnosis, and patients with multiple relapses were not excluded. Doses used were ara-C at 2 g/m2 once daily on days 1–5 and clofarabine at 15, 20, or 25 mg/m² once daily on days 1–5, with dose escalation guided by a “3+3” algorithm; G-CSF 5 μg/kg was given one day before beginning, daily during, and after chemotherapy until recovery of neutrophils.

Results: Sixteen pts have been treated, 9 at 15 mg/m², and 7 at 20 mg/m². Four pts, 2 at each dose level, had received prior allogeneic stem cell transplants; three of the 4 had been transplanted within 5 months of starting G-CLAC, and had relapsed within 2 months after transplant. Grade 3–4 DLTs [pulmonary infection (2), meningitis (1)] were more frequent in previously transplanted pts, occurring in 1/2 pts treated at 15 mg/m² and in 2/2 pts treated at 20 mg/m², and led to exclusion of transplant patients with GVHD or on immunosuppresion from the study. In contrast, the incidence of DLT in evaluable non-transplanted pts was 1/8 at 15 mg/m² and 0/3 at 20 mg/m². The single DLT in the non-transplanted pt was pulmonary toxicity associated with prior history of suspected fungal infection that improved on antifungal therapy. Five of 8 pts without prior transplant have entered CR; the other four pts are too early to evaluate. The lower bound of an exact 95% confidence interval for the CR rate (.24–.91) appears at least comparable to what might be expected with high-dose ara-C (HiDAc), or fludarabine/ara-C/G-CSF (FLAG) regimens for poor risk patients. In particular, only 1 of 16 pts received G-CLAC as 1^st salvage after a first CR >1 year, while the majority had either brief remissions or never achieved CR. Using a prognostic model that derives expected CR rates with HiDAc or FLAG based on 1^st CR duration and number of prior salvage therapies (

Conclusion: Our results suggest that G-CLAC is an effective, well-tolerated regimen in non-transplanted patients with poor prognosis, relapsed or refractory AML.