Does Concomitant Methotrexate (MTX) and GM-CSF Administration after Allogeneic Hematopoietic Stem Cell Transplantation (HSCT) Predispose to Thrombotic Microangiopathy (TMA)?.

Routine use of myeloid growth factors (GF) after allogeneic HSCT from marrow (BMT) or blood (BSCT) results in faster neutrophil recovery but no survival benefit. Registry data suggest that G-CSF administration after allogeneic BMT influences outcome adversely (Ringdén et al. J Clin Oncol2004;22: 416–23). GM-CSF (Anasetti et al. Blood1993;82Suppl 1: 454a) as well as G-CSF (Schriber et al. Blood; 1994:84:1680–4) administration after unrelated donor BMT has been shown to cause higher treatment-related mortality (TRM) and poorer survival. After having noted an apparent increase in TMA after allogeneic HSCT for acute lymphoblastic leukemia (ALL), we reviewed the course of 10 ALL patients (32–53 y, median 40) allografted from HLA-matched siblings in first remission (n=9) or relapse (n=1) after uniform conditioning with 60 mg/kg etoposide and 1320 cGy total-body irradiation (TBI) in 6 fractions. 9 received G-CSF-mobilized BSCT and 1 BMT. 6 were allografted on a study (ECOG 2993/UK MRC ALL 12) that mandated GM-CSF administration from day 0. The other 4 received no GF in accordance with our standard practice of not using myeloid GF post-allograft. All were scheduled to receive cyclosporine, and 15 mg/m2 MTX on day 1 and 10 mg/m2 MTX on days 3, 6 and 11. 8 patients received all 4 doses of MTX, 1 (not on GM-CSF) received 3 doses, and 1 (on GM-CSF) received only the first dose. TMA characterized by elevated LDH (2–12x increase from baseline), schistocytosis, and variable renal dysfunction (1.7–2.7x increase in creatinine from baseline) was seen in 5 patients; all on GM-CSF (5 of 6 GM-CSF recipients vs 0 of 4 not getting GM-CSF; P=0.053). The only GM-CSF recipient not developing TMA had received only 1 dose of MTX. Thus all 5 patients getting GM-CSF and full-dose MTX developed TMA compared with none of the other 5 (P=0.012). The occurrence of TMA was not associated with grossly elevated cyclosporine levels, required cyclosporine cessation in 4 patients all of whom also underwent plasmapheresis, and contributed to death in 2. Because of this safety concern, we have stopped GF administration after allogeneic HSCT even in patients on ECOG 2993. It is unclear whether this toxicity is attributable to GM-CSF through endothelial activation, the combination of GM-CSF and MTX, or the combination of GM-CSF and MTX in this particular clinical setting (ALL treated with high-dose MTX a few weeks before HSCT or high-dose etoposide-TBI conditioning). While MTX is known to be associated with TMA, we have not seen a similar trend in other diseases where patients have received MTX after the allograft but have not received GF. The ECOG-MRC study also may not be able to answer this question readily as a number of centers have used G-CSF instead of GM-CSF or no GF at all. Our data indicate that this association requires retrospective review of patients who have already been treated, and, in combination with other observations, suggest that not only have myeloid GFs failed to improve the outcome of allogeneic HSCT, but safety concerns continue to cloud their routine use in this setting. Additional data are required to clarify the influence of GF use on the outcome of allogeneic HSCT. For the moment, the use of myeloid GFs should probably be confined to patients experiencing delayed engraftment after allogeneic HSCT.