Cyclophosphamide in severe aplastic anemia?

In this issue of Blood, Scheinberg et al investigate moderate-dose cyclophosphamide (120 mg/kg) in treatment-naïve patients with severe aplastic anemia (SAA). The study was stopped for safety reasons.¹ 

SAA is a rare but serious form of bone marrow failure related to an immune-mediated mechanism that results in severe pancytopenia, a high risk of life-threatening infection, and hemorrhage. Allogeneic bone marrow transplantation (BMT) from an HLA-identical sibling donor is the treatment of choice for young patients, leading to an 80% to 90% chance of survival, with no or few complications in the long term.²  In the absence of an HLA-identical sibling donor or in older patients, excellent outcomes can be achieved with the current gold standard first-line immunosuppressive therapy (IST) consisting of antithymocyte globulin (ATG) plus cyclosporine (CsA; 60-70% response rate and 70-80% long-term survival).³  After IST, only one-third of the patients are cured, one-third are dependent on long-term administration of CsA, and one-third will either relapse or develop a clonal disorder (myelodysplastic syndrome or acute myeloid leukemia); the latter complications are usually rare after BMT. Additional immunosuppressive drugs (ie, high-dose corticosteroids, mycophenolate mofetil, or sirolimus) have been added to the ATG/CsA backbone with the objective of decreasing relapse and secondary clonal disease, but thus far, no improvement in outcome has been observed.⁴  The use of rabbit ATG, a more potent immunosuppressant drug than horse ATG, did not improve the rate of hematological recovery and was associated with a detrimental effect on overall survival, justifying the recommendation to prioritize horse ATG in this setting.⁵ 

Cyclophosphamide is as an alternative immunosuppressive drug in treatment-naïve patients with SAA, which may improve outcome but has already generated considerable controversy. A single phase 2 study conducted at the Johns Hopkins University testing high-dose cyclophosphamide (200 mg/kg) was updated in 2010 and an editorial in this journal commented on the update.⁶  The reported high-level range results were of an overall actuarial survival rate of 88% (response rate of 71%) and an actuarial event-free survival rate of 58% in 44 treatment-naïve patients at 10 years.⁷  Unfortunately, this procedure was associated with long-lasting neutropenia and severe fungal infections occurring at a rate of 18.2%. Moreover, in a randomized study conducted at the National Institutes of Health, directly comparing high-dose cyclophosphamide and horse ATG/CsA in treatment-naïve patients, excess toxicity and death from invasive fungal infections were observed in the cyclophosphamide arm, which led to early termination of the protocol.⁸  A Chinese trial of moderate doses of cyclophosphamide (30 mg/kg per day for 4 days) appeared attractive when presented at a specialty conference several years ago and is now available in print,⁹  with an abbreviated period of neutropenia, little morbidity and mortality, and a response rate comparable to rabbit ATG. Those results combined with the recent drastic improvement of antifungal drugs justified further investigation of moderate doses of cyclophosphamide in patients with SAA.

Scheinberg et al report on 22 consecutive treatment-naïve patients with SAA who were treated with moderate-dose cyclophosphamide (120 mg/kg) plus low-dose CsA (recommended therapeutic range between 100 and 200 µg/L). Overall, 9 patients (41%; 95% confidence interval [CI], 20-62%) responded at 6 months (4 complete and 5 partial remissions). The average duration of severe neutropenia (<0.2 × 10⁹/L) was 2 months. The average length of initial inpatient hospitalization was 47 days, and confirmed filamentous fungal infections were documented in 6 cases despite aggressive antifungal prophylaxis. With a median follow-up of 2.2 years, the cumulative incidence of clonal evolution was 22% (95% CI, 0-39%), the cumulative incidence of late events (clonal evolution + relapse) was 28% (95% CI, 4-46%), and the overall survival at 2 years was 72% (95% CI, 55-94%) (see figure). The Data Safety Monitoring Board eventually recommended termination of accrual due to unacceptable toxicity. Of course, differences exist between this study and the Chinese one⁹  that may explain the discrepancies in results (technical issues, ethnic considerations, and daily care). However, 2 (1 randomized) prospective controlled trials of cyclophosphamide at 200 mg/kg or at moderate doses (120 mg/kg) in treatment-naïve patients with SAA are now published, both indicating unacceptable rates of toxicity with no benefit in term of relapse or clonal evolution.

Consequently, the real question is not whether we should continue to use cyclophosphamide in SAA (obviously not outside prospective clinical trials) but whether cyclophosphamide is a good investigational drug for SAA in 2014. Trials in the field are difficult to conduct mainly because of the rarity of the disease and the excellent outcome after standard treatment combining horse ATG and CsA. Moreover, one has to take into account 2 other recent important insights in the field of SAA:

1. The outstanding results of the oral thrombopoietin mimetic eltrombopag in patients with refractory SAA with multilineage clinical responses in some patients. Eltrombopag has just been approved for this indication in the United States by the federal Food and Drug Administration, and preliminary results in previously untreated patients are promising.² 

2. The drastic improvement of unrelated BMT with a recent presentation at the last European group for Blood and Marrow Transplantation (EBMT) meeting from the UK group using this procedure first line (whereas normally reserved to refractory patients after IST failure) and showing excellent results.¹⁰ 

The SAA working party of the EBMT decided to prioritize a prospective randomized trial, which will compare horse ATG plus CsA with or without eltrombopag (http://clinicaltrials.gov/show/NCT02099747) in treatment-naive patients. At this time, unrelated BMT also seems very promising, yet the average 3-month delay between the diagnosis of SAA and BMT, mainly because of donor recruitment, renders such a procedure in patients with severe neutropenia at diagnosis difficult. Thus, cyclophosphamide does not appear to be a clinical research priority in treatment-naïve patients with SAA.