Comorbidity Indices Predict Survival Among Elderly Patients with Acute Myelogenous Leukemia but Do Not Aid in Selecting Between Traditional Induction Chemotherapy and 5-Azacytidine Treatment.

Abstract 1040 Poster Board I-62 Treatment of the elderly patient with acute myelogenous leukemia remains problematic as many have poor performance status precluding aggressive therapy and others have high risk features including abnormal cytogenetics. The success of 5-azacitidine among patients with advanced myelodysplastic syndromes in controlling blast cells and prolonging survival has led to its use in AML. However choosing between potentially curative hospital-based induction chemotherapy and palliative outpatient hypomethylating treatments is difficult. The availability of predictive comorbidity scores may aid in guiding patients in this decision. Methods: Patients with a new diagnosis of AML, ages 60 and older, were offered traditional in-patient induction therapy (7 days cytarabine 100 mg/m2 CI with 3 days idarubicin 12mg/m2) or out-patient therapy (5-azacitidine 75 mg/m2 IV 5 days per month). We report a retrospective review of clinical outcomes among patients treated between June 2003 and August 1, 2009. A review of presenting clinical features, placing patients into prognostic groupings by the Charlson Comorbidity Index (J Chronic Dis 1987; 40:373), Hematopoietic Cell Transplantation – Specific Comorbidity Index (Blood 2007; 110:4606), and the Myelodysplastic Syndrome – Specific Comorbidity Index (Blood 2008; 112: abstr 2677) was performed. Points for a diagnosis of leukemia were excluded from the indices. Results: 99 patients with AML were treated at our center, all with leukemic blasts >20% (median 48%). 75 chose traditional 7+3 and 24 chose 5-aza. The median age was younger for those choosing 7+3 (67 vs 77 years, p<0.001). The median survival for all patients was 340 days (95% CI 258, 422). The median survival was longer with 7+3 at 367 days (95% CI 306,428) compared to a median survival with 5-aza of 186 days (95% CI 122, 250) (p=0.07). Only one patient age <65 chose 5-aza. The median survival for pts with 7+3 was similar by age groups: <65 yrs (n=34) 367 days, 65-69 yrs (n=19) 340 days, and ≥70 yrs (n=22) 467 days (log-rank p=0.31). Cytogenetic features (SWOG criteria) at baseline correlated with survival: good risk (n=3) 186 days, intermediate risk (n=54) 604 days, and poor risk (n=28) 241 days. (p=0.03). Similarly, cytogenetic risk continued to remain significant regardless of treatment choice: 7+3: intermediate/poor risk karyotype 664 days vs 340 days, and 5-aza: intermediate/poor risk 604 days vs 67 days (p=0.03). All 3 examined comorbidity indices were able to predict overall survival for the entire cohort (n=99). Using the CCI, better scores led to improved outcome (log-rank p=0.007): Score 0 (n=41): 475 days; score 1 (n=23): 213 days; score 2 (n=25): 196 days; score 3 (n=8): 186 days; score 4 (n=2) 1404 days. Similar results were obtained using the HCTS-CI (log-rank p=0.01): Score 0 (n=33): 475 days; score 1 (n=24): 353 days; score 2 (n=13): 146 days; score 3 (n=17): 196 days; score 4 (n=5) 1113 days; score 5 (n=5) 337 days; score 8 (n=1) 1404 days. And using the MDS-CI (log-rank p=0.02) better scores predicted improved survival: Score 0 (n=64): 375 days; score 1 (n=18): 241 days; score 2 (n=13): 186 days; score 3 (n=3): 69 days; score 4 (n=1) 1404 days. In addition, patients who scored well on the CCI did better (p=0.005) than patients with poor scores by type of treatment (ie, a low score patient treated with 7+3 did better than a high score patient treated with 7+3, and a low score patient treated with 5-aza did better than a high score patient treated with 5-aza). Similar relationships were noted between scores on the HCTS-CI and MDS-CI and treatment choice. However, in all CCI score categories, 7+3 treatment was superior to 5-aza (score 0: 839 vs 298 days; score 1: 276 vs 69 days; score 2: 279 vs 106 days; score 3: 435 vs 90 days) (p=0.08). Similarly 7+3 treatment appeared superior to 5-aza by HCTS-CI score (p=0.02) and MDS-CI (p=0.05). Conclusions: Elderly patients with AML can achieve survivals approaching one year with current treatment options. 5-azacytidine is a reasonable option for patients declining aggressive hospital based treatments especially if cytogenetic features are favorable, however overall survival may be inferior to standard induction chemotherapy. Although comorbidity indices are able to stratify potential outcomes among elderly patients with AML choosing a particular treatment, in our series we could not identify a comorbidity score that would dissuade aggressive treatment in favor of less intensive therapy. Disclosures: Goldberg: Celgene: Speakers Bureau. Off Label Use: 5-azacytidine use in AML. Masood:Celegene: Speakers Bureau.