We, along with the European Blood and Marrow Transplant Severe Aplastic Anaemia Working Party, have major concerns about the use of high-dose cyclophosphamide without stem cell rescue in patients with aplastic anaemia because of the major risk of early mortality from serious infections (especially systemic fungal infections) as well as bleeding.
The study by Tisdale from the National Institutes of Health (NIH), a centre of excellence for the treatment of aplastic anaemia in the USA, was a prospective randomized study comparing high-dose cyclophosphamide and ciclosporin with the combination of antithymocyte globulin (ATG) and ciclosporin (Tisdale et al, 2000). The study was terminated prematurely, we believe correctly so, on account of three deaths within 3 months from fungal infections and three further cases of proven or suspected fungal infections in the cyclophosphamide arm. There were no early deaths and no fungal infections in the ATG arm. Furthermore, the cyclophosphamide arm was associated with poor response compared with the ATG arm. Data from large multicentre studies conducted since 1990 using ATG and ciclosporin confirm a much lower early mortality risk (Bacigalupo et al, 2000a, b).
The purpose of prospective randomized studies is not only to compare efficacy and outcome but also to perform interim analyses to detect unexpected serious morbidity and mortality as soon as possible. The study by Brodsky et al (2001) and their conclusions illustrate the potential shortcomings of small, uncontrolled studies, particularly for rare diseases, such as aplastic anaemia.
The combination of ATG and ciclosporin represents a well tested and much safer option for treatment. Furthermore, for those patients who fail to respond to ATG and do not have a human leucocyte antigen identical sibling donor, recent results of mini-allografts from matched unrelated donors show significant improvement in outcome (Bacigalupo et al, 2002).
Recent follow-up from the prospective randomized NIH study indicates that high-dose cyclophosphamide does not prevent relapse or clonal evolution in aplastic anaemia and does not eradicate pre-existing paroxysmal nocturnal haemoglobinuria clones (Tisdale et al, 2002). In the absence of any prospective randomized study to show the benefit of high-dose cyclophosphamide in aplastic anaemia, we feel that the results from the NIH study should guide recommendations for the management of aplastic anaemia.