Risk of acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) after adjuvant chemotherapy (CT) for early breast cancer (BC) in the community setting.

Abstract

62 Background: AML and MDS complicate adjuvant CT in BC. Incidence Rates of MDS/AML with pegfilgrastim (PGCSF) use and newer adjuvant regimens in large patient (pt) populations are not widely characterized. Methods: We queried the iKnowMed electronic health record from a large network of community oncology practices for pts diagnosed with stage I-III BC from 2007-2010 with at least 5 visits and follow up (f/u) through 2/2012 for our retrospective study. We stratified pts by adjuvant CT utilization (yes/no), regimen type, PGCSF use, age, and characterized the incidence of MDS/AML captured as a secondary diagnosis. Fisher’s exact test and student t-test were used for categorical and continuous variables, respectively; Cox proportional hazard model was used to estimate hazard ratios (HR) for risk factors associated with AML/MDS development. Results: We identified 20,900 pts with median f/u of 2.8 years (yrs) (1.2-5.2 yrs). 11,295 pts (54%) received CT, 41% of whom received anthracyclines (A); 9,605 (46%) did not receive CT. Median age of diagnosis in the CT and non-CT arms was 54 and 64 yrs, respectively (p < 0.01). Among the CT-treated group, 12 pts or 0.11% (95% CI, 0.06-0.19) developed AML/MDS with median time to onset of 1.8 yrs and median f/u of 2.7 yrs. Of these 12 pts, 8 received A and 11 PGCSF. In the non-CT group, 18 pts or 0.19% (95% CI, 0.11-0.30) developed AML/MDS with median time to onset of 2.2 yrs and median f/u of 3 yrs (p=NS). Multivariate analysis of pts who received CT revealed pts ≥70 vs. <70 yrs and those that received A-containing vs. alternate regimens were more likely to develop AML/MDS. Conclusions: Adjuvant CT did not increase risk of AML/MDS compared with those that did not receive CT. However, our findings confirm that increased age and A-containing CT regimens are associated with increased risk. The low event rate in our study population may be due to short f/u, younger age in the CT treated arm, and high utilization of non-A CT. Association with PGCSF warrants further evaluation. [Table: see text]