Outcomes in relapsed/refractory AML are dismal. Given the high prevalence of multidrug resistance (MDR), many clinical studies have tested whether MDR reversal agents such as cyclosporine (CSA) could increase the efficacy of conventional chemotherapeutics but results have been inconsistent. More recently, HMG-CoA reductase inhibitors (e.g. pravastatin) have been investigated after demonstration that cholesterol synthesis inhibition can restore chemosensitivity of AML cells in vitro. As studies further suggested the HMG-CoA inhibitors could downregulate MDR function, we attempted to improve the efficacy of mitoxantrone and etoposide in patients with relapsed/persistent AML by combining CSA and pravastatin in a single-arm, open-label phase 1/2 trial.
Adult patients with relapsed/refractory non-APL AML and a treatment-related mortality score of <9.2 (Walter et al. J Clin Oncol 2011) were eligible if they had an ECOG performance status (PS) ≤3 and adequate organ function. Prior hematopoietic cell transplantation (HCT) was permissible if relapse occurred >180 days post-HCT. Planned treatment included induction and consolidation therapy with pravastatin, CSA, and increasing doses of mitoxantrone and etoposide. All patients received pravastatin 320mg PO every 6 hours on days 1–10. The starting dose level (level 1) used etoposide 60mg/m2/day and mitoxantrone 5mg/m2/day via continuous infusion on days 5–9. CSA started 6 hours before the first doses of mitoxantrone/etoposide; after loading with 6mg/kg over 2 hours and 4mg/kg over 6 hours, CSA was infused at 18mg/kg/day on days 5–9. CSA levels were monitored on Day 6 and 8 to maintain levels <2,400ng/mL (by LCMS-MS). At dose level 2, identical doses of pravastatin and CSA were combined with etoposide (80mg/m2/day) and mitoxantrone (6mg/m2/day). Dose-limiting toxicities (DLTs) were defined as: 1) any Grade 3 non-hematologic toxicity lasting >48 hours except of febrile neutropenia (FN), infection, or hyperbilirubinemia; and 2) any Grade ≥4 non-hematologic toxicity except FN/infection, constitutional symptoms if recovery to Grade ≤2 within 14 days, and hyperbilirubinemia. An “adaptive” Bayesian phase 1–2 design that monitored both toxicity and response was used. The maximal acceptable rate of toxicity was 30%, the minimum acceptable rate of efficacy 20% (historical rate of 15% in this population), stopping when posterior probability >90%. A maximum of 45 patients would be treated in cohorts of 3.
Between October 2010 and February 2012, 10 patients (median age 52 [range 40–58] years) were treated, including 4 patients (all at dose level 1) who received therapy off protocol. Two patients had secondary AML, and 6 had complex cytogenetics. PS was 1 (7 patients) or 2 (3 patients). Average TRM score was 4.32 for on-study patients. All patients had received at least 2 prior regimens, and 7 received 3 prior therapies. One patient had undergone HCT. The average CSA level on Day 6 was 1,366.3ng/mL (range 713–2,082ng/mL) and 1,199.8ng/mL on Day 8 (range 526–1,593ng/mL). Efficacy (CR or CRi)/DLTs for on-study patients were: level 1 (n=3): 0/1 (grade 4 acute kidney injury); level 2 (n=3): 0/2 (grade 4 acute kidney injury; sepsis/multiorgan failure). Among the 7 patients treated at level 1, other toxicities included septic shock (n=1), FN (4), mucositis (2), and decreased left ventricular ejection fraction (1). One off-study patient achieved CR after 1 course and received 1 cycle of consolidation with pravastatin, etoposide, and mitoxantrone but experienced CNS relapse after CR duration of only 46 days. Based on the results of the on-study patients, the statistical monitoring mandated early study closure for lack of efficacy & excessive toxicity.
In the targeted, heavily pretreated patient population, the study regimen was excessively toxic and failed to achieve acceptable efficacy. Further clinical testing, if any, should be limited to less heavily pretreated, otherwise “fit” patients with AML. This trial was registered at ClinicalTrials.gov as NCT01342887.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.