Response: Further analysis of the relationship between pretransplantation serum ferritin and transplantation outcome: cause-of-death analysis and impact of disease duration

We appreciate the insightful comments of Platzbecker and colleagues on the possible confounders of the relationship between serum ferritin and transplantation outcomes, and on the conclusions that can be drawn from our analysis. At their suggestion, we analyzed the impact of elevated pretransplantation serum ferritin on cause of death, focusing in particular on infection-related deaths (including viral, bacterial, and fungal causes) and pulmonary-related deaths (including diffuse alveolar hemorrhage [DAH], idiopathic pneumonia syndrome [IPS], and respiratory failure). Patients with an elevated ferritin (in the upper quartile) showed a trend toward more infectious deaths in univariate and multivariate analysis (multivariate odds ratio = 1.9, P = .18). The numbers were too small to allow a distinction between pathogen classes. In contrast, there was no evidence of an increased likelihood of dying of pulmonary causes (multivariate OR = 1.0, P = .9). It should be remembered that cause of death is an unreliable variable, particularly in retrospective studies, and we would therefore urge caution when interpreting these results.

In our study, we did not have access to duration of disease-related (pretransplantation) neutropenia, and hence cannot comment on its relationship to ferritin levels. We repeated our multivariable analysis incorporating time to transplantation for the patients with acute leukemia or myelodysplastic syndrome (MDS). In neither case did time to transplantation have an effect on posttransplantation survival (HR = 1.0, P = .4 for patients with acute myeloid leukemia [AML]/acute lymphoblastic leukemia [ALL], and HR = 1.0, P = .3 for patients with MDS). Moreover, the inclusion of this variable in the model had no effect on the hazard ratio associated with an elevated pretransplantation ferritin. Finally, one could argue that in patients with acute leukemia, time to transplantation is very closely related to disease stage (which we accounted for in our models). Therefore, we still submit that pretransplantation serum ferritin in our analysis reflects iron overload rather than disease stage.

We agree with Platzbecker and colleagues that, in the case of patients with low-risk MDS, earlier transplantation may diminish the impact of iron overload on transplantation outcome. However, it is worth pointing out that the decision analysis of Cutler et al¹  used survival figures that presumably incorporate the effect of iron overload on patients who undergo transplantation in later stages of MDS. Moreover, if chelation therapy could be safely performed and could mitigate or even abrogate the detrimental impact of iron overload on transplantation outcome, the opposite may in fact happen: patients could undergo transplantation in later stages of MDS, which could improve quality-adjusted survival.