This is an analysis of 509 patients with severe aplastic anaemia (SAA) treated in Europe between 1981 and 1986; 218 patients were treated by allogeneic bone marrow transplantation (BMT) from HLA identical sibling donors and 291 with immunosuppressive therapy (IS) with antilymphocyte globulin (ALG). The overall actuarial survival was 63% after BMT and 61% after IS therapy at 6 years. All patients fulfilled the criteria of SAA; however, most patients with a neutrophil count of less than 0.2 x 10(9)/l also had infections and haemorrhages. Therefore a further subclassification was defined by pretreatment peripheral blood neutrophil count: very severe aplastic anaemia (vSAA) (less than 0.2 x 10(9)/l neutrophils) and moderately severe aplastic anaemia (mSAA) (0.2-0.5 x 10(9)/l neutrophils). A Cox regression analysis showed that the only significant pre-treatment variables were a low neutrophil count (P = 0.001) and increasing age (P = 0.05). Thus it seemed reasonable to analyse survival data after combined stratification for neutrophils (vSAA versus mSAA) and age (cut off at 20 years). BMT was superior to IS in patients with vSAA under 20 years of age (64% v. 38%; P = 0.01). IS was superior to BMT in patients with mSAA aged 20 or more (82% v. 62%; P = 0.002). The two treatments gave comparable results in young patients with mSAA (BMT = 58%, IS = 62%; P = 0.1), and in older patients with vSAA (BMT = 44%, IS = 43%; P = 0.06). Overall 75/218 and 87/291 patients, given BMT or IS respectively, died. The major cause of failure in BMT patients was graft rejection (n = 22) or problems associated with graft-versus-host disease. For ALG patients the major problem was persistence of the aplasia with haemorrhage (n = 32) or infections (n = 46). This study indicates that over 60% of patients with SAA can be successfully treated with either BMT or IS. Overall survival does not differ in the two groups, though significant differences emerge after stratification for severity of the aplasia and age.

Based on this study, alloBMT from a human leukocyte antigen (HLA)-identical sibling donor does not appear to offer any advantage when compared with ABMT. A graft-versus-Hodgkin effect is associated with > or = grade II aGVHD, but its positive effect on relapse is largely offset by its toxicity. In most circumstances, alloBMT cannot be recommended for patients with HD.

A multi-centre retrospective analysis on 117 patients relapsing after bone marrow transplantation (BMT) for acute leukaemia was carried out by the Leukaemia Working Party of the European Group for Bone Marrow Transplantation (E.B.M.T.). Forty-one patients had acute myeloid leukaemia (AML) and 76 had acute lymphoblastic leukaemia (ALL). Relapse occurred between 3 and 30 months after BMT and where investigated the leukaemia was found to have relapsed in recipient cells. In 10 cases the relapse was associated with new cytogenetic abnormalities. 74 patients received further treatment for leukaemia. Of these 21 out of 50 with ALL and 11 out of 24 with AML achieved a complete remission and had a median survival of 12 months compared with a median survival of 4 months for untreated patients or patients not achieving complete remission (P less than 0.001). Factors predictive for successful remission induction were a long interval between bone marrow transplant and relapse in ALL patients; and isolated extramedullary relapse. Presenting blast count, karyotype and remission status and number at the time of BMT were not predictive. Donor bone marrow was shown to be responsible for haemopoietic recovery occurring in the 21 out of 31 patients tested who achieved remission using donor karyotype or red blood cell antigens as markers. Nine patients received a second bone marrow transplant but only one became a long-term survivor. The results show that chemotherapy can usually prolong survival in selected patients with acute leukaemia relapsing after BMT but further BMT has a poor outlook.

This study suggests that ABMT and alloBMT for non-Hodgkin's lymphoma are comparable, with the exception of lymphoblastic lymphoma in which a graft-versus-lymphoma effect may account for the lower relapse rate for patients who underwent alloBMT.

The definition of resistance is in the eye of the beholderWe appreciate the observations made by Duyster et al that relate to sensitivity of clinical alleles of FIP1L1-PDGFRA associated with hypereosinophilic syndrome to inhibition with AMN107. Our cellular and biochemical experiments are generally in agreement in showing that the FIP1L1-PDGFRA T674I imatinib mesylate resistance allele is also highly resistant to inhibition with AMN107, with more than 100-fold higher IC 50 than for FIP1L1-PDGFRA. 1 There are several potential explanations for why we observed modest effects on cell growth and tyrosine phosphorylation content of FIP1L1-PDGFRA T674I transformed cells at 1 M AMN107 compared with the results annotated here, but at most there may be a several fold difference in our estimates of cellular IC 50 .Several other points merit consideration. First, we have previously reported that PKC412 is a potent inhibitor of the FIP1L1-PDGFRA T674I allele, 2 with an IC 50 of approximately 100 nM vs approximately 376 nM reported here for AMN107. From this perspective, PKC412 may be as appropriate an alternative for FIP1L1-PDGFRA patients who develop resistance due to acquired T674I as AMN107. In this context, it is also worth noting the importance of including controls for nonspecific or off-target toxicities for tyrosine kinase inhibitors such as AMN107 or PKC412, as they are selective rather than specific. Parental Ba/F3 cells would have been 1 such control in these experiments, but in our view the most compelling controls for nonspecific toxicity or off-target effects are kinase mutations that are selected for resistance to the kinase inhibitor in question. For example, we generated a PKC412 resistance allele in the context of FIP1L1-PDGFRA with an N659D substitution. 2 We observed that cells stably transduced with FIP1L1-PDGFRA N659D were resistant to the effects of PKC412, establishing FIP1L1-PDGFRA as the critical target for cellular cytotoxicity mediated by PKC412 in this context, rather than off-target or nonspecific effects. Indeed, the finding of the FIP1L1-PDGFRA T674I imatinib mesylate resistance mutation in a patient with clinically resistant HES provides the most compelling evidence that FIP1L1-PDGFRA causes HES and is the target of imatinib mesylate. It would be useful to have comparable controls for AMN107 as well, although it seems most likely that the cytotoxic effects of AMN107 treatment are attributable to inhibition of the T674I allele.Second, we completely agree that relative resistance of FIP1L1-PDGFRA T674I to AMN107 does not exclude potential clinical applicability and efficacy if adequate plasma concentrations can be achieved. It may be of value to consider either AMN107 or PKC412 as second-line agents. For the time being, the point is moot, as neither of these drugs is Food and Drug Administration (FDA) approved for any indication, but could potentially be accessed through compassionate use mechanisms.It may be, as suggested by the authors, that FIP1L1-PDGFRA T674I will emerge as an important...