Mortality due to conditioning regimen-related toxicity (RRT) has decreased over the past decade, but still remains a major limitation for autologous and allogeneic BMT. In addition, morbidity due to moderate or severe but non-fatal toxicity has a major impact on increased length of stay, cost and therapeutic intervention to treat toxicity while decreasing acute quality-of-life for post-BMT patients. Clinical prognostic factors for treatment-related mortality (TRM) have been described in various reports, however these account for only 40–70% of the variability in TRM, and most of the published prognostic models have evaluated only toxic mortality, not morbidity. Several reports have found associations with individual SNPs and survival, TRM and hepatic toxicity. We hypothesized that genetic variability affecting activity of enzymes that metabolize and detoxify chemotherapeutic drugs and radiation could explain some of the inter-individual variability in TRM and RRT. The glutathione-S-transferase (GST) isoenzymes respond to oxidative stress by detoxifying the reactive oxygen species produced by many chemotherapeutic drugs and radiation therapy. We retrospectively reviewed 496 patients treated with a first myeloablative autologous or allogeneic BMT from 1/1/1996 to 12/31/2002 at RPCI for moderate to fatal (grade 2–4 Bearman RRT criteria) toxicity in association with GSTM1 and GSTT1 genotype. 290 patients (58% of cohort) had procured pre-BMT samples available in the specimen repository, of which 268 patients (92%) yielded amplifiable DNA. DNA was extracted from thawed blood or marrow samples previously stored in liquid nitrogen, using Gentra Puregene DNA isolation kits according to the manufacturer’s instructions, and genotypes were examined using multiplex PCR and visualized by 1.5% agarose gel electrophoresis. 130/268 (49%) patients were homozygous null for GSTM1, 50/268 (19%) were homozygous null for GSTT1, and 17/268 (6%) were homozygous null for both GSTM1 and GSTT1. GSTM1 and GSTT1 individually did not predict overall survival (OS) in autologous or allogeneic BMT patients. However patients who were homozygous null for both genes had a trend toward increased OS in allogeneic but decreased OS in autologous BMT patients, suggesting a differential response to regimen intensity and/or the immunomodulatory effects of graft-versus-host disease. 152/268 patients (57%) developed grade 2–4 RRT in any one or more organ sites. GSTM1 deletion was associated with a significantly increased risk of overall grade 2–4 RRT (RR=1.7, 95%CI 1.02, 2.7; p=0.04), with the same trend seen in both autologous and allogeneic BMT subgroups. GSTT1 deletion alone or in combination with GSTM1 null was not significantly associated with grade 2–4 RRT. Individual organ toxicity was also assessed: GSTM1 deletion was associated with an increased risk of grade 2–4 Stomatitis RRT (RR=1.6, 95%CI 0.99, 2.6; p=0.056); GSTT1 deletion was associated with an increased risk of grade 2–4 GI RRT (RR=3.7, 95% CI 0.96, 14.3; p=0.058). Genotype analysis of SNPs in additional metabolic enzyme gene pathways is ongoing, as well as additional multivariate statistical analyses. The ability to predict moderate to severe RRT after myeloablative autologous and/or allogeneic BMT may improve outcomes by allowing for individualized conditioning regimens.

Disclosures: American Cancer Society.

Supported by American Cancer Society grant # MRSG-05-198-01-CCE.

Author notes


Corresponding author