Purpose: To determine whether comorbidity indexes predict survival in myelodysplastic syndrome (MDS) patients.
Methods: In an IRB approved protocol, we reviewed the records of patients (pts) diagnosed with MDS at the VA New Jersey Health Care System in East Orange, New Jersey, from June 1998 to December 2009. Records were reviewed for demographic, clinical, and pathological data; Eastern Cooperative Oncology Group performance status; erythropoietin use and response; transfusion dependency; International Prognostic Scoring System (IPSS); total number of treatments; and survival. A Cox survival regression analysis was performed. Comorbidity was assessed with 3 comorbidity indexes: the Charlson Comorbidity Index (CMI), the Kaplan-Feinstein Index (KFI), and the Cumulative Illness Rating Scale (CIRS).
Results: There were 81 analyzable points with a median age of 74.5 (patients were aged 47-94). The median hemoglobin was 9.5 g/dL (4.3 to 16.9); median white blood cell count was 5.05 K/cmm (1.2-100); median platelets, 158.5 K/cmm (10-1346); median albumin, 3.9 g/dL (0-5); median ferritin, 378 ng/mL (0-7750); and median LDH, 188 IU/L (0-2426). The median CMI was 2 (0-8); median KFI 2 (0-3); median CIRS 15 4 (0-10); median CIRS 16 7 (0-19); median CIRS 17 1.7 (0-3); and median survival, 925 days (14 to 3871 days). Of the 63 patients, 77.8% received treatment, 34 (42.5%) received an erythroid stimulating agent, 9 (11.25%) lenalidomide, and 8 (10.1%) azacytidine. There were 28 patients (34.2%) who were transfusion dependent with the meidan number of transfusions of 4 (0-100). In the univariate survival analysis, hemoglobin, white blood cells, platelets, ferritin, LDH, albumin, and transfusion dependency were significant predictors of survival. Of the comorbidity indexes, only Charlson CMI and CIRS 19 were significant; age, race, KFI, CIRS 15, 16, 17, and 18 were not significant. In the multivariate analysis, hemoglobin (P < .0040), LDH (P < .0016), transfusion dependency (P < .0028), and CIRS 19 (P < .0303) were independent predictors of survival.
Conclusions: Severe comorbidity, as reflected in the CIRS19, may be an independent predictor of survival. Further analysis of a larger sample will be needed.
